Geoffrey S. Hope and Simon G. Haberle
There are no data on the history of the southern eucalypt savannah woodlands which remain cryptic, since occupation there is only known from the mid-Holocene. Early occupation and the transfer of fauna to the islands of New Britain and New Ireland and the Admiralties by 30–20,000 years ago (O’Connor & Chappell 2004) is not mirrored by evidence of substantial environmental change. Local clearance (in a parallel to Kosipe) has been found at 24,000 BP at Yombon, in inland New Britain (Pavlides & Gosden 1994).
The difference between the southern and northern coasts of New Guinea is that the former has an extensive shelf while the latter is steep. Thus sea level change saw much more dramatic changes to the southern coast (Chappell this volume). An exception to this is the large river systems of the Ramu and Sepik (and to a lesser extent the Mamberamo) which had cut down more than 100 m below their present mouths during the time of lower sea levels. Widespread mangrove remains and shell tools, hundreds of kilometres up river (Swadling & Hope 1992), indicate that a large estuary existed for a few thousand years before it silted up to the present backswamp complex. Although valleys were not incised, similar change has occurred along the south coast as the shores established about 6000 years ago. Rapid siltation and coastal progradation has been demonstrated along the southern coast of Papua (Ellison 2005).
The arrival and effects of Austronesian-speaking agriculturalists Period III ca 4000–Present
The Austronesian arrival around 4000 years ago shows very little correlation with environmental change. Some indications (but almost no local evidence) are that the minor climatic fluctuations known as the little ice ages probably commenced after 3500 BP with small variations in ice extent on the highest mountains. These alternations of slight cooling and warming probably did not change the forest at lower altitudes. However the frequency and severity of El Niño events may have increased, ushering in the drought and frost events that have an effect on cropping and societies at all altitudes (Brookfield 1989; Brookfield & Allen 1989). These droughts may also have allowed fire to extend grasslands into humid forests. Grasslands of Imperata cylindrica (kunai) occur in many areas on poor soils such as the iron- and magnesium-rich ultramafics of Sentani, Telefomin and Popondetta. These clearings are probably of considerable antiquity as regeneration may take thousands of years. In New Ireland and the Jimi Valley, forest clearance within the last 3000 years leads to a short-term phase of gardens for perhaps a few centuries. The sites are abandoned and the mineral soil buried by the buildup of peat swamp due to increased runoff from open vegetation into the valley bottoms (for example Gorecki & Gillieson 1989).
The adjustments of the coast to the establishment of high sea level at 6000 BP continues throughout the late Holocene. For example, beach ridges in New Ireland contain sequences of pottery that are stranded up to 800 m inland by coastal progradation (White et al. 1991). At this time villages probably moved frequently. This earlier dynamic coastline of beach building was replaced by a static lagoonal strand as the growth of coral reefs caught up with the sea level rise and blocked wave action. The modern reef flat is largely dead as it has reached mean sea level, and its growth is concentrated on the seaward edge. This process has been widespread except where abundant river sediment or sand movement has hindered coral growth. Thus for many coastal villages the marine resources available have changed radically in the last 5000 years, with open coasts being replaced by active reefs and finally lower reef productivity.