Polynesian colonization and landscape changes on Mo’orea,French Polynesia: The Lake Temae pollen record

Abstract

An analysis of sediment cores from Lake Temae utilizing pollen, accelerator mass spectrometry (AMS) dating, magnetic susceptibility measurements, and charcoal particle counts was undertaken to assess landscape transformation following Polynesian colonization of Mo’orea in the Society Islands. A significant influx of terrigenous sediment accompanied by increases in charcoal and the presence of the Polynesian cultigen Colocasia (taro) are indicative of human presence on the island by at least 1060–980 cal. yr BP. Polynesian arrival resulted in the rapid alteration of lowland vegetation illustrated in the pollen record by the removal of the coastal tree Pandanus and the promotion of more economically important trees such as Cocos. The most significant period of burning in the charcoal data overlaps with archaeological evidence for expansion into the island interior and the establishment and growth of more intense agricultural practices from 700 to 500 cal. yr BP. The pollen record also documents the apparent abandonment of the coastal plain near Lake Temae during this phase of inland expansion with the terrestrial landscape returning to an environment similar to that found before colonization. A final phase of environmental transformation commenced with European contact at the end of the 18th century and is marked most clearly in the Lake Temae sequence with the conversion of the system to a freshwater body accompanied by the expansion of the freshwater reed Typha and a level of burning not seen at any time over the previous 1000 years. Importantly, the new data from Lake Temae support the colonization model for central Eastern Polynesia of ca. 1000 cal. yr BP and refute the late settlement scenario for the Society Islands.

Keywords

charcoal fire landscape change pollen Polynesian colonization Society Islands

Introduction

Mo’orea, which along with its close neighbor Tahiti comprise the Windward Isles of the Society Islands, has been perhaps the most intensively archaeologically investigated island of this archipelago (Emory, 1933; Green et al., 1967; Kahn, 2005, 2006, 2011; Kahn and Kirch, 2011, 2013; Lepofsky, 1994; Lepofsky and Kahn, 2011; Lepofsky et al., 1992, 1996; Sharp et al., 2010). Nevertheless, significant gaps in our knowledge of the island’s prehistory persist: these include the approximate age of the first arrival of Polynesian colonizers, as well as the sequence and chronology of environmental transformations resulting from centuries of human land use. In order to address both these problems, and as part of a larger research program designed to compare vulnerability and resilience in three contrastive island socio-ecosystems, we undertook to core and analyze the sediments of Lake Temae, on the island’s northeastern coast. Previous sedimentary and palynological investigation of Lake Temae by Parkes (1994, 1997) and Parkes and Flenley (1990) had demonstrated the potential of this site to provide an archive of pollen and other evidence that would bear on the questions of the timing of Polynesian arrival and subsequent environmental transformations.

Initial stratigraphic excavations and radiocarbon dating in the Society Islands, during the 1960s and 1970s, led to an ‘orthodox scenario’ of Eastern Polynesian settlement in which the Marquesas Islands were the first to be colonized, ca. 1650 cal. yr BP, with the Society Islands being secondarily colonized around 1350 cal. yr BP (Jennings, 1979: Figure 1.1; Emory and Sinoto, 1964; Emory, 1979: 219; Kirch, 1986). The discovery of ‘an early form of cultivated coconut’ from anaerobic deposits in the ‘Opunohu Valley of Mo’orea Island, radiocarbon dated to 1360 ± 60 and to 1270 ± 60 BP, led Lepofsky et al. (1992) to claim that Mo’orea had been settled by 1300 cal. yr BP. Over the past 25 years, however, extensive re-dating of early settlement sites in central Eastern Polynesia using the AMS technique on identified, short-lived plant materials has generally failed to replicate the older dates obtained by initial researchers, leading to a somewhat shortened chronology (Allen, 2014; Anderson and Sinoto, 2002; Mulrooney et al., 2011; Spriggs and Anderson, 1993; Wilmshurst et al., 2011). Moreover, the validity of the Mo’orea coconut dates as a proxy for human arrival has been called into question by the recent discovery that coconuts have been naturally present on the island for at least the past 4600 years, and probably longer (Kahn et al., 2015).

Kahn and Sinoto (2017), based on a review of radiocarbon dates from the earliest known occupation sites in the Society Islands, argue that initial Polynesian settlement occurred around 1000 cal. yr BP. The Kahn and Sinoto review was based on over 17 years of re-analysis of Society Islands archaeological sites potentially dating to the settlement period. Anderson and Sinoto (2002) re-dated the Vaito’otia-Fa’ahia sites (ScH1-1, 2) on Huahine which had material culture assemblages with artifacts diagnostic of Archaic East Polynesian culture. The Vaito’otia-Fa’ahia sites originally produced radiocarbon determinations extending to ca. 1150 cal. yr BP (Sinoto and McCoy, 1975). The work by Anderson and Sinoto (2002) built a new site chronology dating to between 900 and 500 cal. yr BP; however, only a few of their samples were derived from short- to medium-lived species. Other dates from these sites have not been reported in full but appear to derive from short-lived materials, with calibrated age ranges listed as a minimum of 920–670 cal. yr BP (Wilmshurst et al., 2011: Table S1). Kahn (2012) has since provided new data from a Mo’orean coastal site (GS-1) found along Cook’s Bay. A wood charcoal sample that might have some in-built age provided a range of 920–670 cal. yr BP. Finally, Kahn and Sinoto’s re-analysis of other Mo’orea coastal sites (2017), including ScMf-5, originally excavated by Green et al. in the 1960s, yielded early colonization dates. A short-lived sample from an earth oven in the deepest cultural deposit yielded a date of 890–690 cal. yr BP (Kahn and Sinoto, 2017) while another coastal site yielded a date of 1150–1050 cal. yr BP (Kahn, unpublished data). Thus, recent reviews of potential colonization period sites in the Society Islands argue for a probable colonization date sometime around or just after 1000 cal. yr BP.

The ‘Opunohu Valley on the northern coast of Mo’orea Island has the most robust dataset in the Society archipelago relating to human land use and settlement patterns (Emory, 1933; Green et al., 1967; Kahn, 2005, 2006, 2011; Kahn et al., 2014; Kahn and Kirch, 2011, 2013; Lepofsky, 1994; Lepofsky and Kahn, 2011; Lepofsky et al., 1992, 1996; Sharp et al., 2010). Based on stratigraphic, paleobotanical, and ethnobotanical evidence, Lepofsky et al. (1996) argued that vegetation clearance and burning associated with shifting cultivation in the uplands of the ‘Opunohu Valley watershed led to massive erosion and deposition of sediments on the valley floor. Direct evidence for initial anthropogenic use of the ‘Opunohu Valley alluvial flat (charcoal from a combustion feature in EU5; Lepofsky et al., 1996: Table 4) was dated to 918–685 cal. yr BP. Sometime after 700 cal. yr BP, there was an intensive inland expansion accompanied by the construction of agricultural terrace complexes (Lepofsky, 1994), along with residential sites and ceremonial complexes (marae; Kahn and Kirch, 2013).

Recently, Kahn et al. (2015) undertook a multiproxy study integrating diverse lines of evidence for transformations of the island’s littoral and lowland zones following Polynesian colonization. Using a combination of archaeological and paleoenvironmental data, Kahn et al. (2015) documented lowland landscape evolution over a timespan of several millennia, including significant transformation of littoral and coastal zones that have obscured earlier human activities and led to significant changes in vegetation and other biota. Beginning as early as 500 cal. yr BP (AD 1420–1490), a major phase of sedimentary deposition commenced which can only be attributed to anthropogenic activity. At several sites, between 1.8 and 3.0 m of terrigenous sediment accumulated over narrow calcareous beach ridges along the coastal zone, as a consequence of active slope erosion. This phase of significant erosion and sedimentary deposition along the coasts and valley bottoms is believed to correlate with the period of major inland expansion of Polynesian occupation and intensive agriculture, indicated by the presence of charcoal throughout the sediments, including wood charcoal from several economically important tree species (Lepofsky et al., 1996).

Pollen analyses as a proxy for vegetation change in the Society Islands were first undertaken in 1985 by Annette Parkes and John Flenley (1990; Parkes, 1994, 1997) who completed an analysis of sediment cores from Lake Temae in Mo’orea, a relatively large and brackish coastal lake. The changes thought to represent early Polynesian landscape change were dated to between 960 and 1260 cal. yr BP, broadly in agreement with the then established archaeological sequence. As part of a larger project being directed by Kahn and Kirch, we present an update of the Lake Temae sequence derived from new pollen cores, which includes a higher resolution dating program, a comprehensive fire history, and a strengthened claim for settlement around 1000 cal. yr BP.

Environmental setting

The island of Mo’orea (17°30′S and 149°50′W; Figure 1), roughly triangular in shape, is the surviving remnant of an ancient basaltic caldera (Quanchi, 2003). Approximately 130 km² in area, like the rest of the Society Islands, it lies in the northeast to southeast tradewind belt. The wetter northeast trades dominate from January to April while the drier southeast trades dominate from May until August. Annual precipitation ranges from 1718 mm at the Mo’orea airport adjacent to Lake Temae on the northeast coast up to 2935 mm at 100 m elevation in the upper Paopao Valley and 3213 mm at 83 m elevation in the mid-‘Opunohu Valley (Meteo France, 2014).

Figure 1.

Location map of Mo’orea, Lake Temae, and the sediment core locations from this study and Parkes (1994).

Lake Temae (Figure 1) is the only lake on Mo’orea. Situated on the northeast coastal plain, this brackish water body is separated from the sea by a calcareous beach ridge that formed on the barrier reef, with the lake basin consisting of the original lagoon. The lake is approximately 1000 m long and 400 m wide, with a maximum depth of 11.3 m (Parkes, 1997: 185). The present-day lake has no direct connection to the sea; however, a 1928 map of the island by Crossland depicts Lake Temae formerly as a much larger water body linked to the ocean via two smaller lakes (cited in Parkes, 1997). Construction of the Mo’orea airport and landing strip obliterated Lake Varea to the southeast, while Lake Motuiti, to the 7 northwest, was already dry in 1985 when Parkes completed her survey and is now beneath a golf course (Parkes, 1997).

Vegetation along the Temae coastline is dominated by cosmopolitan Pacific littoral taxa (e.g. Cordia subcordata, Calophyllum inophyllum, and Heliotropium foertherianum) and common economic and ornamental plants (e.g. Cocos nucifera, Artocarpus altilis, and Hibiscus rosa-sinensis). Coconut groves and Pandanus trees grow around the shores of Lake Temae, while the aquatic introduced reed Typha domingensis dominates the lake margins. Typha is considered a ‘modern’ or European introduction and is thought to have arrived ca. AD 1830 (Meyer, 2016). Several specimens were collected in the 1920s and 1930s and it is now widely naturalized in low elevation wetlands in the Society Islands (Tahiti, Mo’orea, Maiao, Huahine, Ra’iātea, Tahaa; Meyer, 2016). Importantly, it was not seen, reported, or collected in the Society Islands by J. R. Forster during James Cook’s second expedition (1772–1775), although he did cite Typha orientalis in New Zealand (Nicholson and Fosberg, 2004). On Mo’orea, Typha is restricted to the Lake Temae shoreline (Meyer, 2016).

The Temae Valley rises dramatically behind Lake Temae to around 740 m a.s.l., with the valley slopes mostly supporting many dense and historically introduced taxa such as Leucaena leucocephala (Fabaceae; faux acacia), Crotalaria pallida (Fabaceae; rattlebox plant), Tabebuia heterophylla (Bignoniaceae), the grasses Miscanthus floridulus (a’ejo) and Rhynchelytrum repens (natal grass) (Jean-Yves Meyer, personal communication). In addition to the weedy introductions, there is a very rare species of native vine, Colubrina asiatica (tutu) (Rhamnaceae).

In general, the island’s native forest is confined to the higher ridges and peaks; Mo’orea valley interiors and lower slopes are covered in secondary growth that is a mixture of Polynesian and European-introduced species (Meyer and Florence, 1996). By contrast, during the pre-European period (pre-1767), particularly during the 17th–18th centuries at the height of Polynesian occupation and land use, much of the island’s interior was under intensive cultivation, as described in early European and later missionary accounts, the first Tahitian dictionary, and archaeological and paleoecological studies of the ‘Opunohu Valley (Beaglehole, 1962; Bligh, 1792; Corney, 1915; Davies, 1851; De Bougainville, 1772; Ellis, 1829; Forster, 1777, 1778; Kahn et al., 2014, 2015; Lepofsky, 2003; Lepofsky et al., 1996; Parkinson, 1773).

Methods

Three sediment cores from Lake Temae were collected by Athens in 2011 utilizing a Geo-Core system (Colinvaux et al., 1999). The coring locations are shown in Figure 1. Core 1, taken in 6.4 m of water, yielded a 17.4-m-long sediment record. Core 2 and Core 3 are adjacent cores and were recovered in 7.4 m of water, having lengths of 12.37 and 12.0 m, respectively. By way of comparison, Parkes’ (1997: 187–188) core, recovered in 11.3 m of water, was 14.5 m long. Core 3 was selected for sediment, pollen, charcoal, and magnetic susceptibility analyses and shipped intact (in unopened sampling tubes) to the Department of Archaeology and Natural History (ANH) at the Australian National University (ANU) for processing and analysis. Sediment descriptions were carried out for Core 1 and Core 2 at the International Archaeological Research Institute, Inc., in Honolulu, and Core 1 was additionally analyzed for magnetic susceptibility with a Minikappa KLF-3 meter. In Australia, magnetic susceptibility profiles were produced for Core 3 using a Bartington MS2 meter and MSC2 core logging loop. Core 3 was then split into half lengthwise and a sedimentary description compiled.

Sampling for pollen and charcoal analysis was carried out for Core 3 only. For pollen, 2.5 cm³ samples were taken every 20 cm. Based on sedimentary observations and Parkes’ earlier work (1997), the sampling resolution was increased to every 10 cm between 540 and 650 cm, and then every 2 cm from 650 to 690 cm.

Sample preparation for pollen followed standard techniques, with chemical preparations selected to initially disperse sedimentary materials and then progressively remove humic acids, calcium carbonates, bulk organics, cellulose, and silicates, as well as to render the pollen wall ornamentation more visible (Na4P2O7, KOH, HCL, acetolysis, heavy liquid, HF; Bennett and Willis, 2001). A Lycopodium exotic spike was added to determine relative concentrations of pollen and micro-charcoal particles. Pollen identification has been aided by the creation of a pollen reference atlas for Mo’orea based on the ‘Inventaire National du patrimoine Naturel’ administered by the National Museum of Natural History in Paris (MNHN – https://inpn.mnhn.fr/accueil/presentation-inpn) and cross referenced with the ANH pollen reference collection. It should be noted that not all species are represented in the pollen reference collection, but there was a greater than 70% match at genus level.

A target count of between 100 and 200 terrestrial pollen grains and spores was set and attained in all but two samples at 330 and 650 cm. A contributing factor to the very low concentrations in some samples was most likely a result of rapid sedimentation rates.

Charcoal as a proxy for fire was explored in two ways; first, through the quantification of charcoal fragments on pollen slides and second, by the counting of macroscopic charcoal obtained by wet sieving. Macroscopic or sieved charcoal records provide a better representation of fires that occurred locally, besides providing a more continuous temporal sequence than charcoal records from pollen slides alone. However, both microscopic and macroscopic records produce comparable results in terms of broad-scale regional histories of fire (Conedera et al., 2009). Samples for macro-charcoal were taken every centimeter (i.e. continuous sampling) for entire depth of Core 3 (1100 samples). Samples were bleached (sodium hypochlorite) and then sieved through a 125-µm metal mesh. All charcoal particles retained on the 125-µm sieve were counted using a stereo microscope. The charcoal fraction encountered on the pollen slides, micro-charcoal, was counted during the pollen analysis stage using the following criteria: black, opaque, angular particles, and >10 µm. These values were converted to a concentration per cubic centimeter utilizing the Lycopodium spike (see Bennett and Willis, 2001 for formula).

Organic-rich layers were targeted for accelerator mass spectrometry (AMS) 14C dating with 18 determinations obtained for the upper 9 m of core. Bulk samples were submitted and pre-treated by the respective laboratories following standard protocols for such sediments. The age determinations were then calibrated in the program Calib 7.1 using the Southern Hemisphere calibration curve (Reimer et al., 2013). A Bayesian statistics-based age–depth model was developed in the program Bacon 2.2 (Blaauw and Christen, 2011) in R (R Core Team, 2013). The age–depth modeling constructed by Bacon uses Bayesian statistics to reconstruct Bayesian accumulation histories for sediment deposits, through a combination of radiocarbon and other site information (Blaauw and Christen, 2011).

Results

Stratigraphy

Our analysis suggests that there are five major depositional units (DUs) recorded in the four Lake Temae sediment cores; these include the three cores collected in 2011 and our interpretation of the sediment description for the original core collected by Parkes in 1983. The DUs (Figure 2 and Table 1) account for the most significant changes in the four sediment cores, emphasizing coherent cross basin processes.

Figure 2.

Generalized stratigraphic diagram of all three cores collected by this study and by Parkes (1994). Refer to Figure 1 for locations.

Table 1.

Depositional units identified in the Lake Temae cores.

DU	Description
I	Black or very dark brown organic silt; soft and spongy; ca. 1% very fine non-calcareous sand; non-reactive to HCl; unit becomes slightly firmer with depth; dense gastropod lens (Tarebia granifera) in the lower part of DU-I in Core 1 and Core 2; diffuse lower boundary. Parkes’ 1985 core (Parkes and Flenley, 1990) has ‘mineral mud with algal laminations’; sparse plant remains; no marine clastics/shells.
II	Gray silt, spongy texture; firmer than DU-I; no sand in matrix but several calcareous sand lenses are present, one between 6 and 10 cm thick and the other 1 cm thick; thin banding/laminae of darker sediments common in Cores 1, 2, and 3; slightly to non-reactive to HCl; diffuse lower boundary.
III	Light gray silty clay with ca. 2–10% fine calcareous sand; reactive to HCL; dispersed tiny pieces of shell and other marine detritus; limited dark banding near top; Cores 2 and 3 have multiple thin calcareous sand lenses; lower boundary clear.
IVa	Light brownish gray to dark grayish brown silt/silty clay; no sand or marine material; non-reactive to HCl; lower boundary abrupt.
IVb	Dark gray to dark brown silt loam interspersed with a layer of coarse organics and terrigenous sand (Core 1 has discernible plant remains); visible charcoal flecks in Core 3; lower boundary abrupt.
V	Light gray/brownish gray silt loam; occasional layers of calcareous sand; slight increase in shell fragments, but more prevalent in Parkes’ 1985 core; base of cores, lower boundary not observed.

DU: depositional unit.

DU-V, the deepest unit in the 2011 cores, represents a quiescent lagoon setting protected by a calcareous beach ridge to the east, which formed on top of the barrier reef (Parkes, 1997). Ocean water would have circulated as a result of tidal fluctuation throughout the lagoon, entering and exiting through reef channels to the north and south. The build-up of the sand berm may be a product of the earlier highstand drawdown, which generated sand and rubble as the reef edifice eroded to its present level (Dickinson, 2009). Fine calcareous sand was likely carried into the lake by trade and storm winds and wave overwash. Parkes’ 1985 core had several distinctive features, including the prevalence of marine mollusks and barnacles in the silt loam matrix and also lenses of ‘brown mineral mud’ at 9.85–9.91 and 11.25–11.60 m (Parkes, 1994: Appendix 8). Although these minerogenic units were not recognized in Cores 1, 2, or 3 at a corresponding depth, the magnetic susceptibility signature of Core 1 hints at some similarity in deposition (see below).

DU-IV, divided into sub-sections ‘a’ and ‘b’, is striking because of its terrigenous character bracketed above and below by marine sediments. The mostly dark brown silt loam of DU-IVb is a mixture of minerogenic sediment and coarse organics/plant material. Core 2 has a sand lens in the middle, while organics/plant material was confined to the basal centimeter of DU-IV in Core 1. In addition, macro-charcoal flecking was observed in Core 3. DU-IV in Core 3 had occasional small gravel and large sand, suggesting a relatively high-energy event. Because of DU-IV’s depth below sea level, there is no possibility that the water level in Lake Temae ever declined sufficiently such that a mire or wetland formed in the lake basin, and it has the appearance of a significant inwash event. DU-IVa, a grayish brown to dark grayish brown silty clay, appears to be a mixture of lagoonal sediment and minerogenic silt.

DU-III is a gray or light olive gray silty clay with ca. 2% or more fine calcareous sand and seems to be a continuation of DU-V, indicative of ongoing circulation of ocean water through the Lake Temae lagoon. Calcareous sand lenses and occasional fine organic bands of various thicknesses are also present.

DU-II is gray silt with a ‘spongy’ texture and thin dark banding or laminations. The spongy texture and banding/laminations are likely the result of a high proportion of detrital algal remains in the sediment. They suggest reduced tidal circulation and eutrophic conditions. Parkes (1997) suggests that ‘isolation of Lac Temae from the sea may have resulted in an increase in water temperature …’ as well as ‘… an increase in the nutrient status of the lake, [which] may have encouraged the growth of the red and yellow algae/bacteria which dominate the sediments between 2.85 and 1.30 m’. The reduction of ocean water circulation in DU-II may, at least in part, be related to coastal progradation in the areas of the reef channels, thereby restricting water flow through the channels, as well as possible coral growth inside the lagoon and around the channels in these areas.

DU-I is mostly a highly organic silt with a significant minerogenic component from terrestrial sources. This soft, muddy unit represents a continuation of the lake’s eutrophication process and appears to be entirely historic. The uppermost meter of sediment in the location of Core 2 and Core 3 is so soft that it could not be retrieved with the coring device. A dense 10-cm-thick lens of the fresh/brackish water gastropod, Tarebia granifera, a European introduction, was observed in the lower part of DU-I in Core 1 and Core 2 (Carl Christensen, personal communication) and just above the middle in Parkes’ 1985 core (see Parkes, 1994). Its absence in Core 3 is likely the result of a failure to capture the entire drive from this section of the core (see Figure 2).

In summary, all four cores exhibit the same basic stratigraphic patterns, although Parkes’ 1985 core appears to have had a greater marine influence with more shell and sand in DU-V. Also, Core 1 did not contain the numerous mostly thin calcareous sand lenses of the other cores. Furthermore, DU-IVb in Core 1 did not contain the quantity of coarse peat/macro-botanical remains as found in the other cores, suggesting it may have been situated in a more peripheral location from the point of input of this material. Finally, the minerogenic intervals in DU-V of Parkes’ 1985 core were not visually or texturally recognized in the other cores, although Core 1 does appear to have magnetic susceptibility signatures for them.

Magnetic susceptibility

The magnetic susceptibility measurements carried out by Parkes (1997) and those by this study record a major peak in susceptibility associated with the coarse brown mineral layer (DU-IVa and DU-IVb) that occurs in all three of the 2011 cores (Figure 2). The values in Core 3 begin to rise at around 735 cm, reaching a peak in readings from 690 to 660 cm, and then stay relatively high until around 435 cm (Figure 3). Similarly, high readings can be seen in the Core 1 curve and in the core analyzed by Parkes (1997).

Figure 3.

Magnetic susceptibility profiles of all three cores collected by this study as well as by Parkes (1994). The extreme values associated with the terrigenous layer at ca. 680 cm as well as the uppermost sediments of Core 1 have been removed for greater clarity.

Parkes (1997) also carried out geochemical analyses on the Temae sediments and found that this coarse brown mineral layer associated with the peak susceptibility readings is associated with significant values in basaltic oxides, confirming the layer has a terrigenous origin. She also commented on two smaller but significant peaks at around 1160–1118 and 991–965 cm in the 1985 core. Geochemical and physical characteristics suggested that these earlier peaks are a mixture of both marine and terrestrial sediments. While there are similar peaks in Core 1 at 1360, 1160, and 930 cm, subsidiary peaks in Core 3 below 700 cm are much harder to discern (Figure 2).

The uppermost meter of sediment for Core 3 was too soft to collect, but in Core 1 this upper meter has the highest susceptibility measurements in the record by several orders of magnitude. It is suspected that the high values may be because of the accumulation of fine particulates in this region of the lake following the construction of the runway into Lake Temae during the late 1960s and its subsequent extension during the late 1980s (Figure 1). It could also reflect the proximity of LT-1 to the mouth of the Temae River, discharging an abundance of terrestrial silts into the water during the historic period.

Chronology

The 18 AMS 14C determinations from the upper 9 m of Core 3 are reported in Table 2, and the Bacon age model is illustrated in Figure 4. The sediment descriptions already illustrate that sedimentation processes within Lake Temae have varied greatly over time and this is further supported by the dating program. Bacon 2.2 is able to construct a robust age model, even where there are outlying dates, as these are accommodated by use of a Student’s t distribution with wide tails (Christen and Pérez, 2009).

Table 2.

Age determinations (AMS) for Core 3 and the original conventional radiocarbon dates obtained by Parkes (1997); for calibration in years before present (cal. BP) 0 = 1950; age ranges are rounded to the nearest decade.

Lab no.	Sample depth (cm)	Radiocarbon age1σ	δ(¹³C) per mil	Calibrated ageyears BP1σ (median)	Material dated
D-AMS 003944	120–121	Modern	−38.0	Modern	Bulk organic sediment
D-AMS 003945	220–221	632 ± 28	−20.3	630 (605) 550	Bulk organic sediment
D-AMS 005421	319–320	640 ± 27	−36.5	630 (610) 550	Bulk organic sediment
D-AMS 003947	420–421	888 ± 28	−14.8	770 (750) 730	Bulk organic sediment
D-AMS 003948	520–521	720 ± 30	−31.9	660 (640) 570	Bulk organic sediment
Beta-328797	586–587	850 ± 30	−25.0	730 (710) 680	Bulk organic sediment
Beta-411988	614–615	1030 ± 30	−22.4	930 (870) 820	Bulk organic sediment
D-AMS 005422	619–620	1096 ± 28	−21.0	930 (950) 960	Bulk organic sediment
D-AMS 003949	620–621	1126 ± 27	−31.3	1050 (980) 940	Bulk organic sediment
D-AMS 003950	661–662	1303 ± 25	−26.8	1270 (1210) 1120	Bulk organic sediment
D-AMS 005423	661–662	1320 ± 27	−17.6	1270 (1220) 1180	Bulk organic sediment
Beta-411989	665–666	1170 ±30	−25.3	1060 (1020) 980	Bulk organic sediment
Beta-328798	681–682	1090 ± 30	−25.6	960 (940) 930	Bulk organic sediment
Beta-328799	687–688	1170 ± 30	−25.9	1060 (1020) 980	Bulk organic sediment
Beta-411990	712–713	1180 ± 30	−21.3	1060 (1020) 980	Bulk organic sediment
D-AMS 003951	764–765	983 ± 30	−26.2	910 (850) 800	Plant remain
D-AMS 005424	764–765	1118 ± 26	−31.4	1040 (970) 930	Bulk organic sediment
Beta-328800	945–946	2050 ± 30	−25.3	2000 (1960) 1930	Bulk organic sediment
Parkes (1994)
SRR-3088	680–690	1210 ± 90		1180 (1080) 980	Bulk sediment
SRR-3288	1145–1156	1540 ± 100		1510 (1400) 1300	Bulk sediment

Figure 4.

Age–depth model produced by Bacon 2.2 (Blaauw and Christen, 2011).

Several age reversals are evident in the Core 3 sequence, in particular at 420 cm (D-AMS-003947) and at 661 cm (D-AMS-003950 and D-AMS-005423). The δ13C value for D-AMS-003947 at 420 cm (within the calcareous mud) is quite enriched, suggesting a marine influence, possibly contributing to the apparent older age as a result of 14C-depleted water being metabolized by algae. This is further supported by the geochemical analyses carried out by Parkes (1997), which established that sediments below 2 m contain, on average, between 30% and 40% CaO. The exception to this are the two distinct regions of terrestrial input from where her dating samples were taken – less than 20% at 11.5 m and less than 2% at 6.85 m. It is likely that all age determinations, except those taken from within the distinct terrigenous layers, are affected by old marine carbon.

One aim of this study was to more comprehensively date the terrigenous layer from 664 to 682 cm. Parkes dated a 10-cm bulk sample from within the dark brown mud just above the coarse mineral layer, which calibrates to 980–1180 cal. yr BP with a median probability of 1080 cal. yr BP (Table 2; SRR-3088; 680–690 cm). As a comparison, three 1-cm interval samples from the same sedimentary region within Core 3 were dated: 687–688 cm from the dark brown mud that precedes the coarse mineral layer, 681–682 cm from within the coarse highly organic mineral layer, and 665–666 cm within the overlying dark brown mud. The three samples have median age calibrations of 1020, 940, and 1020 cal. yr BP, respectively (Table 2), in agreement with Parkes’ original date. Of note is that the highly terrigenous sediment at 681–682 cm is slightly younger than the bracketing sediment. With four overlapping age determinations, however, we can have considerable confidence that these dating results concerning the terrigenous dominated sediment layers provide a secure anchor for our chronological model based on Bacon 2.2, with the age model placing this change in sedimentation to between 1060 and 990 cal. yr BP (Figure 4). As may be seen in Figure 4, even if there is some concern about the accuracy of the non-terrigenous layer dates because of unaccounted variations in 14C-depleted carbon in the sediment samples, the graph shows only relatively minor age deviations for these samples. We therefore believe the age model provides relatively accurate results even if we do not have an ideal level of precision.

Above 520 cm (ca. 800 cal. yr BP), sediments appear to be dramatically affected by old carbon input until the system becomes essentially freshwater at ca. 200 cm (ca. 310 cal. yr BP). The average sedimentation rate calculated by the age model above the terrigenous layer is 0.66 cm/yr, while the average sedimentation rate below this is somewhat slower at 0.45 cm/yr.

Pollen and charcoal analysis

Because of the rapid sedimentation rate within lake, in combination with the dominance of entomophilous (insect) pollination for much of the tropical flora, pollen concentrations are extremely low in the Lake Temae record, averaging around 10,000 grains/cm³. The pollen is primarily derived from the lake fringing vegetation as well as the surrounding lowland forest. While the pollen record comprises over 70 discriminated pollen types, it is dominated by just four – Pandanus, Cocos, Ficus, and fern spores – with Typha (bulrush) overwhelming all these pollen types in the most recent part of the record, overlapping with European colonization (Figure 5). All these four taxa grow in the immediate vicinity of the lake. As described earlier, Typha on Mo’orea is considered a European introduction; however, this study and that of Parkes (1994) reveal that small quantities of Typha pollen are found in the Lake Temae sediments as early as 940 cal. yr BP. The possibility that Typha was initially introduced by Polynesian people therefore cannot be ruled out.

Figure 5.

Core 3 percentage pollen diagram plotted against depth with calibrated ages shown. Pollen sum does not include Typha given its overwhelming presence in the uppermost samples. Typha is expressed as a percent of the total pollen sum. Charcoal influx: black triangles ⩽1000 pieces/cm³; Typha: green triangles ⩽1%. Economically important taxa and other lowland taxa are shown as triangles as the majority of percentage values are less than 1%. Gray shading in Zone 2 = terrigenous unit; gray shading in Zone 5 = clay layer.

What is not contentious is the introduction of the important economic plant Colocasia esculenta (taro) by Polynesian colonizers, with small quantities of these pollen grains found in sediment layers between 686 and 676 cm (1060–1000 cal. yr BP). The highest counts of this pollen type, however, occur at around 150 cm (ca. 220 cal. yr BP).

All these findings are consistent with Parkes’ original analyses (1997). A comparison of the Pandanus curve from this study with that of Parkes (1997), along with the charcoal records from each study and the presence of Colocasia, is shown in Figure 6 to illustrate the coherency between the respective studies.

Figure 6.

Comparison figure of Pandanus, Colocasia, and charcoal from Core 3 and Parkes (1994). Charcoal: numerical values = truncated values, black crosses ⩽1000 pieces/cm²; solid line = Parkes (1994) terrigenous unit; dashed line = Core 3 terrigenous unit.

Also consistent with Parkes’ original analysis is the dinoflagellate record. Although also found in freshwater, the bulk of dinoflagellates are marine plankton. In the Lake Temae record, they are common in the basal calcareous sediments, more or less disappearing from the sequence after 500 cal. yr BP (Figure 5). This transition, starting at around 780 cal. yr BP, is mirrored by an increase in Cyperaceae (sedge) pollen, signally a shift from a more open marine system to a more closed and terrestrial environment.

Parkes (1997) identified six zones based on the geochemistry of the sediments and changes in the pollen assemblage. Zonation of the current pollen record was carried out by running an optimal splitting by sum of squares analysis on the stratigraphically constrained data within PSIMPOLL (Bennett, 2005). This analysis also resulted in six pollen zones. The following describes the pollen diagram (Figure 5) based on the identified zones accompanied by approximate age ranges in brackets derived from the age model in calibrated years BP, where present equals AD 1950 (Blaauw and Christen, 2011).

Zone 1: Below 700 cm (>1060 cal. yr BP)

The pollen assemblage in this zone is dominated by Pandanus, Cocos, and fern spores. The common coastal tree Barringtonia is also present and there are abundant dinoflagellates and other marine cysts. Pandanus (screw-palm) is often the dominant vegetation above high water mark on many Pacific islands and is an important economic plant to many Polynesian communities. The pollen signature is consistent with the macro-botanical evidence described by Kahn et al. (2015) from anaerobic deposits in the coastal lowlands that predate Polynesian colonization, with abundant preserved drupes of Pandanus and endocarps of Cocos.

There is some uncertainty as to whether Casuarina, an important economic timber tree, is a Polynesian introduction in Tahiti, although it is considered to be a Polynesian marker in other places (e.g. Prebble and Wilmshurst, 2009). If it is a Polynesian introduction, it appears in Core 3 of the Lake Temae record for the first time at around 1070 cal. yr BP. Parkes (1997) found Casuarina in an isolated sample from the base of her core, but more consistently in her record in the upper 2.5 m (ca. the past 400 cal. yr BP).

Fire in the landscape is also apparent in this zone, revealing that fire was a natural part of the ecosystem prior to human arrival (Figures 5 and 6). The macro-charcoal peak in this zone (see Figure 6) may represent a local or catchment level fire; however, there is no obvious disruption in the vegetation record.*

Zone 2: 700–600 cm (~1060–910 cal. yr BP)

The sample resolution in this zone is much higher than for the rest of the record and it encompasses the minerogenic layer (DU-IVa) thought to represent a significant inwash event. Significant changes in the pollen assemblage for this zone include the dramatic decline in Pandanus and Cocos percentages and the increase in fern spore values. Most importantly, however, the distinctive echinate pollen grains of Colocasia (taro) appear for the first time, with isolated grains found at depths of 688, 684, 676, 620, and 630 cm (Figure 5). The earliest appearance of Colocasia at 687–688 cm, and hence the earliest evidence of human presence in Core 3, has an AMS determination of 1060–980 cal. yr BP.

Trilete fern spores rise significantly in this zone; the most common type is Dicranopteris. Dicranopteris fernlands in Mo’orea are thought to be maintained through periodic burning and are regarded as an indicator of disturbed or strongly nutrient-depleted soils, with the micro-charcoal record suggesting that fire is more prevalent after this time (Figure 5). Other notable changes in the pollen assemblage include the appearance of Poaceae (grass) and an increase in the disturbance indicators Trema and Urticaceae. Ficus, a common lowland tree, obtains values of around 10%, and several other lowland forest trees are also present, such as Glochidion, Ixora, Macaranga, and Psychotria type. Both Glochidion and Macaranga are common lowland taxa that can also be regarded as pioneer taxa and good indicators of disturbance or secondary forest. In a similar way, many Ficus species can be rapid colonizers after disturbance (Whittaker et al., 1989). Several other taxa that are of economic importance, apart from Colocasia, such as Casuarina, Hibiscus, and Mangifera, either appear for the first time or have a more sustained presence in the record from this time onward. This is also the zone where Typha pollen appears for the first time.

The pollen assemblage of Zone 2 appears to hold several lines of evidence for human presence in, and transformation of, the Mo’orea landscape: the appearance of Colocasia, the removal of the common lowland trees (Pandanus and Cocos), the rise of secondary tree taxa and ferns associated with disturbance, as well as the macro- and micro-charcoal records suggesting burning has become more regular and sustained in the landscape. The age range of this zone and interpretation of human presence are consistent with the initial colonization timeframe now proposed for Mo’orea and other central Eastern Polynesian islands (Conte and Molle, 2014; Kahn and Sinoto, 2017; Kirch et al., 2010).

Zone 3: 600–420 cm (~910–690 cal. yr BP)

Values of Pandanus fall away dramatically and are replaced by Cocos, an important lowland economic tree for Polynesian people. Ficus pollen disappears as does the pollen of most of other lowland tree taxa found in the previous zone. Some of the highest levels of macro-charcoal (local burning) outside of the historic period occur in this zone, while micro-charcoal (more of a regional signal) remains unchanged (Figure 6).

The three dinoflagellate types as well as the other marine indicators are either absent from the outset of this zone or have disappeared by the end. The decline in marine microfossils plus the change in sedimentary character at this time suggest that Lake Temae was slowly being cut off from direct contact with the ocean. Cyperaceae becomes the dominant fringing plant, further evidence of a substantial substrate change. Small amounts of Typha pollen continue to be found in this zone.

It would appear that throughout this zone Tahitians have continued to modify the landscape, removing taxa that have little economic importance while encouraging the growth of economically important trees such as Cocos. At the same time, inadvertent sedimentation within the lagoon system may have been cutting Temae off from ocean circulation as reflected in the sediment characteristics and the loss of marine indicators. This would be in keeping with the scenario of vegetation clearance and burning associated with shifting cultivation in the uplands leading to erosion and deposition on the valley floors (Lepofsky et al., 1996).

Zone 4: 420–260 cm (~690–425 cal. yr BP)

This pollen assemblage records a dramatic decline in the economically important Cocos which is offset by an increase in Pandanus. In addition, the lagoon fringing sedges (Cyperaceae) decline. Pioneer taxa such as Ficus once again become more prevalent in this zone as does Glochidion. The marine indicators are absent and the sediment is primarily algal. The highest values of micro-charcoal (regional burning) outside of the historic period occur during this period accompanied by reduction in macro-charcoal (local burning; Figure 6).

It would appear that Lake Temae at this point is now a brackish and eutrophic water body and as a consequence Tahitians may have abandoned this locality for more desirable habitat, leading to the expansion of secondary tree taxa and Pandanus around the shoreline. This would be in keeping with the archaeological evidence for an intensive inland expansion accompanied by the construction of agricultural terrace, residential sites, and ceremonial complexes (marae) sometime after 700 cal. yr BP (Kahn and Kirch, 2013; Lepofsky, 1994). It is worth noting that there is limited archaeological work from around Lake Temae and none from the Temae valley.

Zone 5: 260–160 cm (~425–240 cal. yr BP)

This section of the record is defined by declining Pandanus values and increasing representation of Cocos, Cyperaceae, and ferns. Cocos, however, does not attain the values it had prior to 720 cal. BP. The presence of Urticaceae pollen suggests that some level of disturbance was occurring in the surrounding landscape; however, all burning as reflected in the charcoal records is subdued, suggesting a period of stability with regard to human activity.

Zone 6: 160–100 cm (<240 cal. yr BP)

The final zone in the Core 3 record covers a potential age range of 300 cal. yr BP (AD 1650) to 90 cal. yr BP (AD 1860). However, it appears that the later part of this zone (and DU-I) is missing because of lack of recovery of the sediments, including the Tarebia granifera gastropod lens. Thus, the dating of this zone in Core 3 probably starts within the first one or two decades following initial European contact and extends through the early- to mid-19th century.

The most notable change in this zone is the shift to a freshwater environment, based on the characteristics of the sediment, as well as the pollen signature which is overwhelmed by Typha (bulrush), an obligate freshwater wetland species. Typha domingensis today occupies all the mudflats around Lake Temae with this expansion commencing at around 150 cm (AD 1730–1785) and overlapping with initial European contact ca. AD 1770. This Typha expansion is also in keeping with the findings of Parkes’ (1997) original work. However, as noted earlier, rather than this being a signal purely of a European introduction, it seems more likely to be the creation of suitable habitat, as Typha pollen is found in small quantities in earlier levels.

The highest values of charcoal in both size fractions are also recorded in this zone, with the values of uppermost samples two and three times those seen in earlier samples. This is in keeping with ideas that the use of fire has intensified during the historical period across the Pacific. The zone is also notable for having the highest absolute counts of Colocasia (taro) pollen also at 150 cm overlapping with the start of European contact, as well as the highest values of Casuarina pollen, an important economic tree species.

Parkes recorded a number of 19th- and 20th-century introduced weeds in the uppermost meter of her 1985 core. As noted earlier, however, sediment from the most recent historical period could not be retrieved at the Core 3 location, as it was too loose for retention in the sample tube; these taxa are therefore missing from this record. What is undoubtedly captured by this record, however, is the commencement of European landscape transformation in Mo’orea.

Charcoal

Charcoal as a proxy of fire is found throughout the record, with both the macro- and micro-charcoal records revealing charcoal deposition well before ca. 1000 cal. yr BP and thus a natural element of this environment (Figures 5 and 6). In both records, however, there is an increase and more sustained input of charcoal after 688 cm (ca. 1010–1040 cal. yr BP).

As noted in the previous section, the most pronounced period of regional burning, as represented by the smaller charcoal fraction counted on the pollen slides, is in Zone 4, estimated to cover the period from ca. 690 to 425 cal. yr BP, and again in the most recent part of the record after European contact. The earlier period overlaps with the archaeological models of community settlement and agricultural expansion utilizing slash and burn to transform wild landscapes into anthropogenic ones (Lepofsky et al., 1996; Kahn and Kirch, 2013; Kahn et al., 2015). There are several peaks of macro-charcoal throughout Zones 3 and 4, representing local fire, with the largest quantity of this fraction found at ca. 830 cal. yr BP. Local fire throughout the record corresponds with the conspicuous loss of lowland taxa of less economic importance and with the increase in taxa of greater economic importance, such as Cocos, suggesting a managed landscape.

Parkes (1997) also determined charcoal content as a proxy of fire within the Lake Temae sediments and also found evidence of burning well before Polynesian arrival (Figure 6). However, the methodology employed by Parkes was somewhat different, utilizing nitric acid digestion of organic matter followed by combustion, and as far as can be determined, the varying content of organic matter down the sediment core was not used to correct the charcoal curve and could explain some of the divergences from this study. This would be especially true in the less organic marine sediments below around 700 cm.

Discussion

The sedimentary record from Lake Temae largely captures changing hydrological conditions within the system, including the varying inputs from marine and terrestrial sources. Several of the original analyses undertaken by Parkes (1994, 1997) have also been undertaken here, with replication of the pollen record, age determinations, and mineral magnetic measurements illustrating a high level of coherency between the two studies.

Lake Temae is now only a remnant of a once larger coral reef lagoon that occupied the entire northeastern coast of Mo’orea (Figure 1). The deepest sediments recovered have the appearance of being laid down within an enclosed barrier reef, which in all likelihood was a low-energy environment with a regular inflow and outflow of ocean water through narrow channels in the barrier reef – intermittent storm surf overwash bringing in coarser marine sediments. The sediments that preceded Polynesian colonization of the island are dominated by marine deposition, while those post colonization appear to have less of an open marine influence, suggesting channel restriction that essentially caused the lake to become the brackish water body seen today (Parkes, 1994, 1997).

The original dating program carried out by Parkes (1994, 1997) had two age determinations on samples that were largely terrestrial in origin. Even though a more extensive dating program has been undertaken by our study, an obstacle we could not overcome is the lack of suitable dating material for much of the core (that is a lack of macro-botanical remains) along with uncertainties about the proportion of old carbon in the calcareous muds. However, the coarse minerogenic layer found in all cores recovered from Lake Temae at an average depth of 6.8 m below sediment surface and found to be almost entirely terrigenous (Parkes, 1994, 1997) has a dated age range in Core 3 of 1060–950 cal. yr BP (Table 1). This age range overlaps with that determined by Parkes (1997: 188; Table 1) and is in close agreement with the archaeologically estimated age range for Polynesian colonization of ca. 1150–1050 cal. yr BP (Kahn and Sinoto, 2017; Lepofsky and Kahn, 2011), while being slightly older than the timeframe proposed by Wilmshurst et al. (2011) of ca. 925–830 cal. yr BP.

The terrigenous layer is also notable for the lack of marine indicators. This suggested to Parkes (1994, 1997) that the earlier minerogenic bands, with their mixture of marine and terrestrial sediments, were most likely the result of storm activity, while the overwhelmingly terrigenous layer around ca. 1050–910 cal. yr BP was most likely the result of land clearance overlapping with a significant climatic event (Parkes, 1994, 1997) A possible scenario is that a brief but very intense episode of storm runoff from the Mo’orea hillslopes and drainage basins west of Lake Temae resulted in terrestrial sediments colluvially and fluvially transported into the lake, depositing coarse organics as well as fine mineral sand found in Cores 2 and 3. The organic material and coarse plant remains found in this layer were no doubt scoured from shoreline wetlands that had formed on the landward side of the lake.

Hillslope erosion, vegetation change, and the creation of new landscapes after colonization by people have been reported from numerous Pacific Islands locations (e.g. Hope and Spriggs, 1982; Kennett et al., 2006; Prebble and Wilmshurst, 2009; Prebble et al., 2013; Spriggs, 1997; Stevenson, 2004; Stevenson et al., 2001) including Mo’orea (Kahn et al., 2015). As noted earlier, the 1928 map by Crossland (cited in Parkes, 1994, 1997) reveals that less than a century ago Lake Temae was much larger and connected to a lagoon system through two smaller lakes, one to the north and one to the south. Parkes (1994, 1997) also suggested that prior to 1000 cal. BP all three lakes were most likely part of the lagoon system, with the coral beach ridge to the east of the lake at that time a coral islet on the barrier reef. The change from a more or less open lagoon system to a more restricted system occurred well after the termination of the mid-Holocene highstand, which Dickinson (2009: 7) estimates occurred ca. 2000 cal. BP in the Society Islands, and well after his estimated crossover date of 1500 cal. BP, when ‘ambient high tide eventually fell below the surfaces of paleoreef flats constructed at mid-Holocene low-tide level’. Dismissing sea level change as a viable possibility, coastal infilling and progradation may be at work, a result of hillslope destabilization by Polynesian settlement and agriculture as suggested by Parkes (1997). This process was likely initiated by the expansion of prehistoric settlement and agriculture utilizing slash and burn practices into the inland valleys as early as 950–750 cal. BP, with well-developed communities archaeologically documented by 550 cal. BP (Green, 1996; Kahn, 2005, 2007, 2011, 2013; Kahn and Kirch, 2011, 2013; Lepofsky, 1994; Sharp et al., 2010). It seems most likely that coastal deposition of terrigenous sediments overwhelmed the old lagoon to the north and south of the current Lake Temae with sediment, thus restricting marine through-flow. It was not until after European contact, however, that Lake Temae became completely cut off, forming the brackish water body we know today, with the small lake to the north now incorporated into the golf course, while the southern lake lies beneath the airport.

The interpretation that terrigenous sedimentation within Lake Temae is associated with Polynesian colonization of the island is also supported by the pollen record. Intermittent and isolated grains of Colocasia (taro), an important Polynesian cultigen, are detected for the first time at ca. 688 cm in Core 3, with the commencement of terrigenous inwashing although preceding the coarse terrigenous layer. An AMS age determination at 688–687 cm calibrates to 1060–980 cal. yr BP. Colocasia pollen grains are not present in the coarse terrigenous layer, although the very low counts of this pollen type in the surrounding sediments are in keeping with the plant being a low pollen producer as well as the likelihood of it being grown well beyond the lake boundaries. It is worth noting that the highest counts of Colocasia in the core occur around the time of European arrival. All the findings with regard to Colocasia are once again in keeping with Parkes’ original work.

Significant catchment disturbance at ca. 1000 cal. yr BP is supported by other elements of vegetation change, namely, a reduction in the taxa that dominated the pollen record up to that point, Pandanus and Cocos, in association with the appearance or increase in secondary/disturbance taxa such as Macaranga, Trema, and Urticaceae and the fern Dicranopteris. Some lowland forest taxa such as Ficus, Glochidion, Ixora, and Hibiscus also become more apparent and grass appears for the first time at around 1050 cal. yr BP. Taken together, the pollen record suggests that lowland vegetation is characterized by disturbance taxa or open patches from this time forward, although the low values of grass pollen in no way suggest the conversion of the landscape to grassland. Overall, however, this would be consistent with the development of slash and burn activities as fire is also sustained after ca. 1050 cal. yr BP.

From ca. 1050 to at least 700 cal. yr BP, the sediment of the lake is a combination of terrestrial and marine sediment. Marine indicators, however, begin to decrease at ca. 900 cal. yr BP and have disappeared altogether by ca. 700 cal. yr BP as marine inflow is thought to become more restricted through siltation. The pollen record suggests that over this time the immediate landscape was more open, with the terrestrial pollen spectrum dominated by Cocos and the shoreline taxa dominated by Cyperaceae. Unlike the earlier assemblage, there is only a minor representation of Pandanus. The expansion of Cocos and the decline in Pandanus from ca. 900–700 cal. yr BP further support the exploitation and opening up of the lowland landscape, in particular the favoring of economically important trees such as Cocos.

This situation is reversed at ca. 700 cal. yr BP, with the almost complete loss of Cocos pollen, the resumption of Pandanus dominance, and an increase in the colonizer Ficus as well as several other secondary tree and disturbance taxa. These changes, however, follow on from a 10-cm pulse of marine sand that most likely represents a major storm event that altered the vegetation and hydrology of this strip of coast, possibly leading to Tahitians moving their agricultural/subsistence activities to the more fertile and protected interior. The micro-charcoal record clearly reflects slash and burn activities at a regional rather than local scale from ca. 700–500 cal. yr BP.

The archaeological record, independent of the pollen record, suggests that this timeframe of 700–500 cal. yr BP is a period of inland expansion involving the construction of agricultural terraces, residential sites, and ceremonial complexes (marae; Kahn and Kirch, 2013; Kahn et al., 2014; Lepofsky, 1994). This expansion intensified after ca. 500 cal. yr BP, resulting in a major phase of erosion and lowland deposition. Evidence for this is found at several locations along the north coast where between 1.8 and 3.0 m of terrigenous sediment containing charcoal of economically important species was deposited over the narrow calcareous beach ridges after 500 cal. yr BP (Kahn et al., 2015). It is most likely that in the Lake Temae sediments a similar period of sedimentation resulted in the silting up and closing off of the northern and southern channels, converting Lake Temae into a stagnant and brackish water body.

The uppermost 2.5 m of sediment captures both pre-European and post European landscapes. At ca. 210–250 cm, there is an influx of gray clay, after which the system becomes predominantly freshwater. Just how this sedimentary event changed the hydrology of the lake is unclear, however; the pulse of clay occurs somewhere between 300 and 400 cal. yr BP, with no significant alteration to the local vegetation. This may be a local expression of the agricultural intensification and resulting sedimentation discussed in the previous paragraph. Europeans first visited Mo’orea ca. AD 1770, and shortly after this time there is an intense shift in the local vegetation. Typha (bulrush) increases dramatically reflecting the presence it has along the lake margins today. Pandanus is removed from the landscape and Cocos returns as does Colocasia (taro) along with the economically important wood Casuarina.

Annette Parkes (1997) captured the ongoing arrival and expansion of European weeds in her 1985 core. Unfortunately, the top meter of sediment was not captured in Core 3, although the Core 1 sediments from this study did reveal that the uppermost meter is defined by further large-scale inwashing of terrestrial sediment. This is no doubt a consequence of European impact, in particular the large-scale works associated with the construction and later expansion of the Mo’orea airport and runway in 1967 and 1987, respectively.

Conclusion

The analysis of sediment cores from Lake Temae, including AMS dating, pollen analysis, magnetic susceptibility, and micro- and macro-charcoal particle counts, provides significant new evidence bearing on (1) the likely date for Polynesian colonization of Mo’orea Island (and by extension, of the Society Islands) and (2) the impact of Polynesian land use on the island’s vegetation. The date of ca. 1300 BP proposed by Lepofsky et al. (1992) for Polynesian settlement of Mo’orea based on the findings of ‘semi-domesticated’ coconuts preserved anaerobically in ‘Opunohu Valley sediments is now considered unlikely, given the documented presence of coconuts on the island in pre-human times (Kahn et al., 2015) as well as the presence of Cocos pollen throughout this core and that of Parkes (1997). The influx of terrigenous sediments in Zone 2 of the Lake Temae sequence, accompanied by increases in charcoal and by the presence of the Polynesian cultigen Colocasia (taro), is indicative of human presence on the island by at least 1060–980 cal. BP. This age range is similar to that proposed by Kahn and Sinoto (2017) of 1150–1050 cal. BP based on radiocarbon dates from archaeological contexts. The earliest pollen core dates for human activity on Mo’orea have age ranges that are consistent with other recent assessments of initial colonization dates for other central Eastern Polynesian islands (Conte and Molle, 2014; Kirch et al., 2010). Holistic archaeological studies allowing for dates on evidence for early cultural activity including plants and animals introduced via colonizing populations argue for initial settlement of central Eastern Polynesia ca. 1050–950 cal. yr BP (Conte and Kirch, 2004; Kirch, 2011; Kirch et al., 2010; Molle, 2011). Molle and Conte’s recent work at Hane (Molle, 2011; see also Anderson et al., 1994) and the work by Kirch et al. at the Onomea site on Mangareva are some of the strongest evidence for settlement in the region dating to ca. 1050–950 cal. yr BP (see Kahn, 2014). All these new data considered together support a rapid regional colonization or ‘advancing wave’ of colonists into central Eastern Polynesia (Allen and McAlister, 2013; Kahn et al., 2015) somewhat contradicting the late settlement scenario for the Society Islands proposed by Wilmshurst et al. (2011).

The Lake Temae sequence also contributes to our emerging understanding of the changes Mo’orea incurred following the arrival of Polynesians as a consequence of their land use practices. Almost immediately following human arrival, there is evidence for the alteration of lowland vegetation, exemplified by the removal of Pandanus and the promotion of more economically important trees such as Cocos. The increased occurrence/use of fire in association with slash and burn cultivation is borne out by the charcoal data, with the intensity of these activities leading to a heightened vulnerability of steep slopes to erosion, supporting the findings of Kahn et al. (2015) for the widespread deposition of interior sediments in the coastal lowlands. The record also documents the apparent abandonment of the coastal plain near Lake Temae during this phase of inland expansion and intensification from 700 to 500 cal. yr BP with the terrestrial landscape returning to an environment similar to that found before colonization. A final phase of environmental transformation commenced with the arrival of Europeans at the end of the 18th century; this is marked most clearly in the Lake Temae sequence with the conversion of the system to a freshwater body, expansion of the freshwater obligate reed Typha, and a level of burning not seen at any time over the previous 1000 years.

Footnotes

Acknowledgements

We thank Alex Morrison, International Archaeological Research Institute, Inc., for his assistance with aspects of the Bacon depth–age model. Carl Christensen provided the Tarebia granifera identifications, while Jean-Yves Meyer Délégation à la recherche, Government of French Polynesia, provided valuable data and discussion of the Mo’orea vegetation. We thank John Flenley and an anonymous referee for their comments and suggestions. Permission to conduct research at Lake Temae was granted by Chantal Tahiata, Ministèrè de la Culture, de L’Artisianat, et de la Famille; Teddy Tehei, Chef de Service, Service de la Culture et du Patrimoine; Priscille Frogier, Chef de Service, Délégation à la Recherche de la Polynésie Française; and M. Van Bastolaer, le Maire de Mo’orea-Maiao. Bellona Mou and Tamara Maric of the Service de la Culture et du Patrimoine facilitated the permit process. Neil Davies, Director, Hinano Murphy, Frank Murphy, and Valentine Brotherson of the University of California Gump Research Station assisted our fieldwork in various ways.

Funding

This project was funded by National Science Foundation collaborative grants BCS-1029765 to JK and BCS-1030049 to PVK.

References

Allen

(2014) Marquesan colonisation chronologies and post-colonisation interaction: Implications for Hawaiian origins and the ‘Marquesan Homeland’ hypothesis. Journal of Pacific Archaeology 5(2): 1–17.

Allen

McAlister

(2013) Early Marquesan settlement patterns of interaction: New insights from Hatiheu Valley, Nuku Hiva Island. Journal of Pacific Archaeology 4: 90–109.

Anderson

Sinoto

(2002) New radiocarbon ages of colonization sites in East Polynesia. Asian Perspectives 41: 242–257.

Anderson

Leach

Smith

et al . (1994) Reconsideration of the Marquesan sequence in East Polynesian prehistory, with particular reference to Hane (MUH1). Archaeology in Oceania 29: 29–52.

Beaglehole

(1962) The Endeavour Journal of Joseph Banks 1768–1771, 2 vols. Sydney: Angus and Robertson.

Bennett

(2005) Documentation for Psimpoll 4.25 and Pscomb 1.03: C Programs for Plotting Pollen Diagrams and Analysing Pollen Data. Uppsala: Department of Earth Sciences, University of Uppsala.

Bennett

Willis

(2001) Pollen. In: Smol

Birks

HJB

Last

(eds) Tracking Environmental Change Using Lake Sediments, Vol 3: Terrestrial, Algal and Siliceous Indicators. Dordrecht: Kluwer, pp. 1–32.

Blaauw

Christen

(2011) Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Analysis 6: 457–474.

Bligh

(1792) A Voyage to the South Sea Undertaken by Command of His Majesty for the Purpose of Conveying the Bread-Fruit Tree to the West Indies, in His Majesty’s Ship the Bounty, Commanded by Lieutenant William Bligh. London: George Nicol.

10.

De Bougainville

(1772) A Voyage Round the World, Performed by Order of His Most Christian Majesty, in the Years 1766, 1767, 1768 and 1769 (trans Forster

J.R.

). London: Printed for J Nourse and T Davies.

11.

Christen

Pérez

(2009) A new robust statistical model for radiocarbon data. Radiocarbon 51(3): 1047–1059.

12.

Colinvaux

De Oliveira

Patiño

JEM

(1999) Amazon Pollen Manual and Atlas. Amsterdam: Harwood Academic (Gordon and Breach).

13.

Conedera

Tinner

Neff

et al . (2009) Reconstructing past fire regimes: methods, applications, and relevance to fire management and conservation. Quaternary Science Reviews 28(5): 555–576.

14.

Conte

Kirch

(eds) (2004) Archaeological Investigations in the Mangareva Islands, French Polynesia (Contributions of the Archaeological Research Facility, No. 62). Berkeley: University of California.

15.

Conte

Molle

(2014) Reinvestigating a key site for Polynesian prehistory: new results from the Hane dune site, Ua Huka (Marquesas). Archaeology in Oceania 49(3): 121–136.

16.

Corney

(1915) The Quest and Occupation of Tahiti by Emissaries of Spain during the Years 1772–1776, Vols. 2 and 3. London: Cambridge University Press.

17.

Davies

(1851) A Tahitian and English Dictionary. Tahiti: London Missionary Society.

18.

Dickinson

(2009) Pacific atoll living: How long already and until when? GSA Today 19(3): 4–10.

19.

Ellis

(1829) Polynesian Researches, 2 Vols. London: Fischer, Son and Jackson.

20.

Emory

(1933) Stone Remains in the Society Islands (Bernice P. Bishop Museum Bulletin 119). Honolulu, HI: Bernice P. Bishop Museum.

21.

Emory

Sinoto

(1964) Eastern Polynesian burials at Maupiti. Journal of the Polynesian Society 75: 143–160.

22.

Forster

(1777) A Voyage Round the World in His Brittanic Majesty’s Sloop, Resolution, Commanded by Captain James Cook, during the Years 1772, 3, 4 and 5, 2 Vols. London: Printed for B White, J Robson, P Elmsly and G Robinson.

23.

Forster

(1778) Observations Made during a Voyage around the World in Physical Geography, Natural History, and Ethic Philosophy. London: G. Robinson.

24.

Green

(1996) Settlement patterns and complex society in the Windward Society Islands: Retrospective commentary from the ‘Opunohu Valley, Mo’orea. In: Julien

Orliac

(eds) Mémoire de Pierre, Mémoire d’Homme: Tradition et Archéologie en Océanie. Paris: Publications de la Sorbonne, pp. 209–228.

25.

Green

Rappaport

et al . (1967) Archeology on the Island of Mo’orea, French Polynesia. Anthropological Papers of the American Museum of Natural History 51(2): 111–230.

26.

Hope

Spriggs

(1982) A preliminary pollen sequence from Aneityum Island, southern Vanuatu. Bulletin of the Indo-Pacific Prehistory Association 2: 88–94.

27.

Jennings

(ed.) (1979) The Prehistory of Polynesia. Cambridge, MA: Harvard University Press.

28.

Kahn

(2005) Household and Community Organization in the Late Prehistoric Society Islands (French Polynesia). PhD Thesis, University of California, Berkeley.

29.

Kahn

(2006) Society Islands (Central Eastern Polynesia) Chronology: 11 Radiocarbon dates for the Late Prehistoric Expansion and proto-historic periods in the ‘Opunohu Valley, Mo’orea. Radiocarbon 48(3): 409–419.

30.

Kahn

(2007) Power and precedence in Ancient House Societies: A case study from the Society Island Chiefdoms (French Polynesia). In: Beck

(ed.) The Durable House: House Society Models in Archaeology (Occasional Paper No. 35). Carbondale: Center for Archaeological Investigations, pp. 198–223.

31.

Kahn

(2011) Multi-phase construction sequences and aggregate site complexes of the Prehistoric Windward Society Islands (French Polynesia). Journal of Island and Coastal Archaeology 6: 24–50.

32.

Kahn

(2012) Coastal occupation at the GS-1 site, Cook’s Bay, Mo’orea, Society Islands. Journal of Pacific Archaeology 3(2): 1–10.

33.

Kahn

(2013) Temple renovations, aggregate marae, and ritual centers: The ScMo-15 Complex, Lower Amehiti District, ‘Opunohu Valley, Mo’orea (Society Islands). Rapa Nui Journal 27(2): 33–49.

34.

Kahn

(2014) Colonization, settlement, and process in Central Eastern Polynesia In: Hunt

Cochrane

(eds) Handbook of Prehistoric Oceania. Oxford: Oxford University Press, pp. 1–11.

35.

Kahn

Kirch

(2011) Monumentality and the materialization of ideology in Central Eastern Polynesia. Archaeology in Oceania 46: 93–104.

36.

Kahn

Kirch

(2013) Residential landscapes and house societies of the late prehistoric Society Islands (French Polynesia). Journal of Pacific Archaeology 4(1): 50–72.

37.

Kahn

Sinoto

(2017) Refining the Society Island cultural sequence: Colonization period and developmental period coastal occupation on Mo’orea Island. Journal of the Polynesian Society 126(1): 33–60.

38.

Kahn

Horrocks

Nieuwoudt

(2014) Agriculture, domestic production, and site function: Micro-fossil analyses and late prehistoric landscapes of the Society Islands. Economic Botany 68(3): 246–263.

39.

Kahn

Nickelsen

Stevenson

et al . (2015) Mid-to Late-Holocene landscape change and anthropogenic transformations on Mo’orea, Society Islands: A multi-proxy approach. The Holocene 25(2): 333–347.

40.

Kennett

Anderson

Prebble

(2006) Prehistoric human impacts on Rapa, French Polynesia. Antiquity 80(308): 340–354.

41.

Kirch

(1986) Rethinking East Polynesian prehistory. Journal of the Polynesian Society 95: 9–40.

42.

Kirch

(2011) When did Polynesian settle Hawai’i? A review of 150 years of scholarly inquiry and a tentative answer. Hawaiian Archaeology: 123–126.

43.

Kirch

Conte

Sharp

et al . (2010) The Onemea site (Taravai Island, Mangareva) and the human colonization of southeastern Polynesia. Archaeology in Oceania 45(2): 66–79.

44.

Lepofsky

(1994) Prehistoric Agricultural Intensification in Society Islands, French Polynesia. PhD Thesis, University of California, Berkeley.

45.

Lepofsky

(2003) The ethnobotany of cultivated plants of the Maohi of the Society Islands. Economic Botany 57(1): 73–92.

46.

Lepofsky

Kahn

(2011) Cultivating an ecological and social balance: Elite demands and commoner knowledge in Ancient Ma’ohi Agriculture, Society Islands. American Anthropologist 113(2): 319–335.

47.

Lepofsky

Harries

Kellum

(1992) Early coconuts on Mo’orea Island, French Polynesia. Journal of the Polynesian Society 101: 299–308.

48.

Lepofsky

Kirch

Lertzman

(1996) Stratigraphic and paleobotanical evidence for prehistoric human-induced environmental disturbance on Mo’orea, French Polynesia. Pacific Science 50(3): 253–273.

49.

Meteo France (2014) Fiche Climatologiques, Afareaitu 2, Paopao 1, Papetoai 2, Papetoai 3.Statistiques 1981–2010. Toulouse: Meteo-France.

50.

Meyer J-Y (2016) Les zones humides de polynesie française : un ecosysteme meconnu, mesestime et menace. Contribution À la Biodiversité de Polynésie Française N°19, Sites Naturels D’intérêt Ecologique et Patrimonial VIII. Territoire de la Polynésie Française: Délégation à la recherche, pp. 23.

51.

Meyer

J-Y

Florence

(1996) Tahiti’s native flora endangered by the invasion of Miconia calvescens DC (Melastomataceae). Journal of Biogeography 23: 775–781.

52.

Molle

(2011) Ua Huka, une île dans l’histoire: Histoire pré- et post-européene d’une Société Marquisienne. PhD Dissertation, University of French Polynesia.

53.

Mulrooney

Bickler

Allen

et al . (2011) High-precision dating of colonization and settlement in East Polynesia. Proceedings of the National Academy of Sciences 108: E192–E194.

54.

Nicholson

Fosberg

(2004) The Forsters and The Botany of the Second Cook Expedition (1772–1775). Regnum Vegetabile, Vol. 139. Ruggell: ARG Gantar Verlag.

55.

Parkes

(1994) Holocene environments and vegetational change on four Polynesian Islands. PhD Thesis, University of Hull.

56.

Parkes

(1997) Environmental change and the impact of Polynesian colonization: Sedimentary records from Central Polynesia. In: Kirch

Hunt

(eds) Historical Ecology in the Pacific Islands: Prehistoric Environmental and Landscape Change. New haven, CT: Yale University Press, pp. 166–199.

57.

Parkes

Flenley

(1990) Hull University Mo’orea Expedition, 1985 (School of Geography and Earth Resources Miscellaneous Series 37). Hull: University of Hull.

58.

Parkinson

(1773) A Journal of a Voyage to the South Seas in His Majesty’s Ship the Endeavour, Faithfully Transcribed from the Papers of the Late Sydney Parkinson, Draughtsman to Sir Joseph Banks, Esq. on His Late Expedition with Dr. Solander, Round the World. London: Stanfield Parkinson.

59.

Prebble

Wilmshurst

(2009) Detecting the initial impact of humans and introduced species on island environments in Remote Oceania using palaeoecology. Biological Invasions 11(7): 1529–1556.

60.

Prebble

Anderson

Kennett

(2013) Forest clearance and agricultural expansion on Rapa, Austral Archipelago, French Polynesia. Holocene 23(2): 179–196.

61.

Quanchi

(2003) Atlas of the Pacific Islands. Milton: John Wiley & Sons.

62.

R Core Team (2013) R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing.

63.

Sharp

Kahn

Polita

et al . (2010) Rapid evolution of ritual architecture in central Polynesia indicated by precise 230Th/U coral dating. Proceedings of the National Academy of Sciences 107(30): 13234–13239.

64.

Sinoto

McCoy

(1975) Report on the preliminary excavation of an early habitation site on Huahine, Society Islands. Journal de la Société des Océanistes 31(47): 143–186.

65.

Spriggs

MJT

(1997) Landscape catastrophe and landscape enhancement: Are either or both true in the Pacific? In: Kirch

Hunt

(eds) Historical Ecology in the Pacific Islands: Prehistoric Environmental and Landscape Change. New Haven, CT: Yale University Press, pp. 80–104.

66.

Spriggs

MJT

Anderson

(1993) Late colonization of East Polynesia. Antiquity 67: 200–217.

67.

Stevenson

(2004) A Late-Holocene record of human impact from the southwest coast of New Caledonia. The Holocene 14(6): 888–898.

68.

Stevenson

Dodson

Prosser

(2001) A late Quaternary record of environmental change and human impact from Plum Swamp, New Caledonia. Palaeogeography, Palaeoclimatology, Palaeoecology 168: 97–123.

69.

Whittaker

Bush

Richards

(1989) Plant recolonization and vegetation succession on the Krakatau Islands, Indonesia. Ecological Monographs 59: 59–123.

70.

Wilmshurst

Hunt

Lipo

et al . (2011) High-precision radiocarbon dating shows recent and rapid initial human colonization of East Polynesia. Proceedings of the National Academy of Sciences 108(5): 1815–1820.