The Sepik–Ramu System
Setting - The Sepik and Ramu catchments drain the northern side of the New Guinea Mobile Belt, with catchment areas of 77,700 km2 and 18, 500 km2, , respectively. Much of the geomorphic description in the following relies on Chappell (1993). The drainage network in each catchment collects in lowland tectonic basins. Unlike the Fly-Strickland system, the Sepik-Ramu drainages have major sediment delivering tributaries along their full length, and these tributaries commonly join the mainstem across low angle fans (Loffler, 1977). Both rivers flow north, crossing low hills and enter the sea about 20 km apart. The uplands area includes Mesozoic geosynclinal and metamorphic rocks, igneous rocks including ultrabasics, and Tertiary sedimentary rocks, and deformed Tertiary terranes. The Finnisterre Range (north side of the Ramu basin), which drains into the Ramu, is raising at 0.8 to 2.1 mm/yr (Abbott et al., 1997). In the main cordillera, uplift rates exceed 0.5 mm/year. Local rates of uplift between 0.5 and 8 mm/year have been observed in the younger ranges north of the Sepik. Rainfall is about 3 to 5 m/year and discharge variability is very low; annual variation is less than a factor of 2 and the 10 year flood exceeds the 2 year floods by only 15 to 20% (Chappell, 1993).
Terrestrial Sediment Delivery - Sediment yield based on limited measurements exceed 1000 tonnes/km2-yr (Chappell, 1993). The sediment source in the highlands is dominated by landslides. Hovius et al. (1998) have argued that drainage network evolution of the Finisterre Mountains is driven by large-scale landsliding.
The Sepik emerges from the central ranges as a braided, gravel bedded river and then shifts to sand where the river turns to the east (Loffler, 1977). The Sepik- Ramu plain (Figure S1) is crossed by the Sepik at a slope of about 1:20,000 and the Ramu with a slope of 1:4000 (Chappell, 1993). Tidal influences reach about 100 km on the Sepik. The rivers are bordered by scroll bar recording channel migration across a 5 to 10 km wide track (somewhat narrower than the lower Middle Fly). The lateral migration rate on the Sepik, however, is much higher than on the Fly and there is an absence of levees along the river. There has also been significant channel shifting on the tributaries; the Yuat River was once about 25 km upstream of its present location. Outside the meander track the Sepik floodplain is about 70 km wide. Migration rate on the Sepik is much higher than on the Fly, reaching values of 30 m/year. Late Quaternary deformation may have caused the Ramu to shift from being a tributary of the Sepik (perhaps along the present coarse of the Keram) to finding an independent path to the coast.
Chappell (1993) summarizes the tectonic development of the Sepik- Ramu basin. He infers that the "tectonic basin" was an inland sea during the Late Pleistocene high sea stands. During the late Holocene as sea level approached a stable level and then a shallow brackish sea was progressively infilled with sediment, perhaps over a 3000 year period. The deltaic plain is a relatively recent development. Freshwater from the two rivers may have drained out through the Bunapas Gap around 6000 BP. Hence, sediment delivery to the coast probably progressively increased through the Holocene as the inland sea filled such that presently relatively little sediment is being lost between the uplands and the sea, (Chappell, 1993).
Submarine Sediment Delivery - The Sepik River, with an estimated annual sediment load of ~100 x 106 t yr-1, empties directly into a submarine canyon that transverses a narrow continental shelf (< 5 km) on the north coast of Papua New Guinea. As such, it serves as a possible analogue for rivers discharging onto margins during low stands of sea level. Recent studies of the Sepik (Kineke et al., in press) suggest a substantial portion of the riverine sediment appears to be transported seaward in the canyon, most of the remainder being stored at least temporarily on the proximal shelf and slope. Salinity and suspended sediment distributions along the axis from the river mouth to the coastal ocean suggest that sediment dispersal is via a plume with both surface and near-bottom components. Rapid sedimentation may occur just seaward of a shallow bar at the head of the canyon, ~ 1 km upstream of the river mouth. The rapid settling probably results from a combination of factors including reduction of turbulent mixing in the presence of strong salinity stratification, convergent bottom flows, and changing particle characteristics. Sediments initially trapped at the bar might continue down the steep slope as a hyperpycnal (negatively buoyant) flow, or be deposited temporarily near the bar and then flow down the canyon as an episodic turbidity current. Shallow seismic observations, along with textural and radioisotope measurements of seabed sediments, reflect the two distinct dispersal pathways. An acoustically transparent (muddy?) drape extends across the outer shelf and slope west of the river mouth, suggesting deposition from a surface plume. In come cases, Holocene muds are seen burying shelf edge reefs. Silt-rich sediments with anomalously low 210Pb activities along the axis of the canyon are consistent with the concept of a near-bottom plume. X-radiographs and radiochemical profiles suggest active canyon processes. Intermediate turbid layers observed along isopycnal surfaces in deeper water suggest a possible terrestrial source for elevated levels of particulate aluminum and iron at depth in the equatorial Pacific. The formation of a divergent plume may be common in other rivers emptying onto a steep slope during flood periods.
Divergent Plume - The processes that cause the vertical divergence of a river plume (e.g., high concentrations of suspended sediment, trapping of sediments at fronts, boundary layer processes, gravity) are not unique to any particular river, therefore vertically divergent plumes may be common to many moderate to large size rivers or rivers of any size which may experience episodic increases in suspended sediment load during a short period of time. For example, an analysis of 150 world rivers showed the importance of small and medium sized rivers in their ability to trigger underflow at their mouth (Mulder and Syvitski, 1995).
The occurrence of hyperpycnal flows may have a major bearing on our understanding of the transfer of material to the ocean basins and might help to explain many discrepancies in our present understanding of river discharge and shelf sedimentation. Examples include discrepancies between river plume suspended sediment budgets and river sediment discharge estimates (e.g., Kineke and Sternberg, 1995) and lack of correlation between surface plume distribution at sea and shelf distribution (e.g., Wheatcroft et al., 1997). On collision margins, the rivers more often discharge particulates and dissolved materials directly to the coastal ocean. Hyperpycnal flows may take more direct routes across the margin and decreased residence times may have a significant impact on such broad issues as carbon cycling, nutrient supply, and primary production on the continental shelf and slope. Furthermore, at lower stands of sea level during the last glacial maximum, many rivers that now have estuaries crossed the continental shelf and emptied into the ocean basin. These are often characterized by submarine canyons (Hudson) or sedimentary deposits at depth (Amazon cone). The Sepik River may provide a modern analogue to river dispersal systems that prevailed during lower stands of sea level.
Sepik-Ramu Selection Criteria:
A. Natural Factors
1. Has strong forcing associated with rapid uplift, earthquakes and landsliding
2. Has active sedimentation with opportunities to study interaction with many kinds of sediment storage elements.
3. Opportunity to investigate documented divergent plume system.
4. Sediment is actively transferred among environments
5. If both the Sepik and Ramu are studied, then it should be possible to create a sediment budget for terrestrial deposition and offshore discharge.
6.There is a high resolution stratigraphic record, especially for the coastal tectonic basin.
B. Human considerations
1. Logistics are challenging because of the lack of transportation infrastructure, but perhaps access through ships coming of the Sepik will make up for this deficiency.
2. Minor anthropogenic influence.
3. Leverage opportunities exist with the TROPICS program and with the PNG government.