Benthic and pelagic communities in the central Gulf of Papua, a major deltaic clinoform located on the south coast of Papua New Guinea, reflect the spatial and seasonal patterns in physical sedimentation conditions and the availability of food resources across this system. The distribution of terrigenous and marine organic substrates, the transport of riverine-derived sediment, and seafloor dynamics are largely controlled by two seasons and wind patterns: the relatively quiescent NW Monsoon and more energetic SE Trades. The physically disturbed inner-topset deposits (<20 m) are dominated at all times by bacterial biomass and have depressed numbers of macrofauna, with evidence of periodic recolonization recorded in sedimentary structures. The inner-topset region also shows strong seasonality in the production and availability of labile organic matter and in bacterial activity. As measured by the concentrations of photopigments and microbial biomass in the upper 2 cm of surface sediments and in the water column, the comparatively quiescent NW Monsoon period is associated with net delivery of reactive substrates as labile organic matter to the bottom and net growth of the benthic microbial community, whereas the Transition period between the NW Monsoonal-Winds (NW Monsoon) and SE Trade-Winds (SE Trades), and the SE Trades, are periods of reactive substrate depletion in bottom deposits and decreased benthic bacterial and macroinfaunal biomass. In contrast to the inner-topset, the less physically disturbed outer-topset and foreset show relatively enhanced, but still low, abundances of large macrobenthos and possible evidence of variation related to food supply. The greatest observed seasonal variation in macrofauna occurs in the Umuda Valley off the Fly River, an apparent mobile sediment conduit.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The tropics dominate the global hydrologic cycle and deliver ∼50–70% of all riverine sedimentary debris, terrestrial carbon, and runoff to the oceans [Milliman and Meade, 1983; Milliman and Syvitski, 1992; Meybeck, 1988; Ludwig et al., 1996; Richey, 2004]. High relief islands of the Indo-Pacific archipelago, such as New Guinea, supply roughly half of the tropical inputs [Milliman and Syvitski, 1992; Milliman et al., 1999]. The coastal ecosystems characterizing the tropical boundary regions are critical components in the processing, modification, and storage of the riverine material fluxes. Benthic and pelagic communities within these ecosystems interact with and reflect physical conditions over a variety of timescales [Alongi, 1998]. With respect to sedimentation processes, the distributions, abundances, and activities of organisms are often strongly impacted by episodic or periodic physical events such as water column overturn, sediment resuspension, and organic matter deposition [Alongi, 1995; Aller and Stupakoff, 1996; Aller and Todorov, 1997]. In the case of subtidal benthos, part of these impacts comes from the coupling of food availability and physical regime, for example, the relative supply of labile planktonic and refractory organic matter to the seafloor during sediment transport events or productivity pulses following water column mixing. Part comes from disturbance of sediment structure and life habit, and part from factors such as direct mortality during burial or erosion. Because biological communities interact with and closely reflect properties of the physical environment and depositional facies, they can be useful indicators of hydrodynamic and transport processes past and present, particularly in remote locales where it is not possible to extensively instrument and continuously monitor sedimentary dynamics.
 In this study, we examine selected benthic and water column biological properties indicative of physical regime and sedimentary facies in the central Gulf of Papua, a major clinoform deltaic system located on the south coast of Papua New Guinea [Walsh et al., 2004]. Previous research in this system has shown that because of energetic sedimentation conditions, bacteria and other microfauna rather than macrofauna typically dominate benthic biomass and activities over large areas of the shallow topset in the Gulf and within inshore estuarine channels of the delta plains [Alongi, 1995; Alongi and Robertson, 1995; Todorov et al., 2000; Aller and Aller, 2004]. Although frequent physical reworking of the bottom generally inhibits development of macrobenthic communities, preserved biogenic sedimentary structures show that some deposits are periodically colonized and extensively bioturbated during relatively quiescent periods with less intense physical reworking or massive sediment deposition [Alongi et al., 1992; Aller and Aller, 2004; Walsh et al., 2004]. In contrast to the macrofaunal community, the shallow topset microbial community is highly active, diverse, and abundant [Alongi, 1995; Todorov et al., 2000; Aller and Aller, 2004]. Food supply also affects community development, and an overall direct correlation between water column primary production and benthic community abundance has been observed both spatially and temporally across the Fly Delta–central Gulf system [Alongi and Robertson, 1995].
 Although a general outline of the biological community patterns has been established, the data are relatively minimal and there are very few coherent seasonal measurements published from the area, particularly in the central Gulf. Our approach in the present study is to relate community characteristics and measures of food resources with known physical sedimentation factors. This is done seasonally at relatively well-studied sites along a central Gulf transect and at new topset sites. Then, we infer the likely behavior of the system more generally during periods and regions for which measurements are not available. In an attempt to reveal longer term trends or lack thereof, we also compare benthic biological data obtained in the present study with those from several temporally sporadic samplings in the central Gulf and proximal Fly River delta over the past 20 years. We demonstrate that regular spatial and seasonal patterns are evident in the distribution and abundance of benthic and pelagic organisms, and in food resources across the major deltaic zones in the Gulf. The relative patterns of dominant bacterial and subordinate macrofaunal biomass appear to be essentially stable over decadal timescales. Macrofaunal abundances and biomass within the foreset can vary substantially between years, perhaps largely in response to allochthonous food supply or grazing pressure. On the basis of the patterns observed in biological, radiochemical, and biogeochemical properties, we propose a general conceptual model for the coupling between physical forcing, food supply, and biological community responses seasonally across the central Gulf clinoform. This model may act as a basis for future systematic biological sampling of the system.
2. Study Area
 The Gulf of Papua lies along the south coast of Papua New Guinea in the northern Coral Sea (8°S, 144°E) and consists of a half-moon shaped shelf area 360 km wide covering >50,000 km2 (Figure 1). A series of rivers drain into the Gulf including from west to east: Fly, Bamu Turama, Omati, Kikori, and Purari. In aggregate these rivers discharge annually ∼365 × 106 metric tonnes of sedimentary particles and 470–690 km3 of water [Salomons and Eagle, 1990; Milliman, 1995; Pickup and Chewings, 1983; Wolanski et al., 1995]. The Fly River is the largest individual system, with a drainage basin covering 76,000 km2. Sediment and water discharge rates from the Fly range from 85–115 × 106 metric tonnes a−1 and 190–220 km3 a−1, respectively. Most sediment accumulates on the inner shelf of the central Gulf in <50 m of water as part of an overall prograding clinoform deposit [Harris et al., 1996; Walsh et al., 2004]; however, off the Fly River, a portion moves seaward within incised relict river channels such as the subaqueous Umuda Valley [Martin et al., 2008].
 The Gulf experiences two principal seasons with three major wind patterns: the NW Monsoons (December–March), the SE Trades (late May–October), and transitional conditions between these periods. In this paper we distinguish the periods from April to late May as Transition (NW Monsoon to SE Trades) and the November period (SE Trades to NW Monsoon) as late SE Trades. Rainfall of 10–13 m a−1 year falls over steep mountainous terrain, varying in source between the NW Monsoonal and SE Trade-Wind periods, and causing landslides and unrelenting large river discharges of water and sediment throughout the year [Pickup and Chewings, 1983; Wolanski et al., 1984]. Major interannual reductions in runoff do occur, however, during El Niño years when significant drought conditions occur over the island [Moi, 2001; Walsh et al., 2004; Ogston et al., 2008].
 The NW Monsoon winds (averaging ∼5.0 m s−1) blow southeastward off the Island resulting in relatively calm sea conditions with wave heights averaging 0.3 m [Thom and Wright, 1983]. The SE Trade winds on the other hand average 5.0–7.5 m s−1, are generally directed north/northwestward [McAlpine et al., 1983] and produce waves averaging 1.3 m [Thom and Wright, 1983]. As a result, disturbance of the seabed at water depths <50 m is more frequent during the trade wind period, when sediment particles are resuspended and kept in suspension as much as 60–80% of the time compared with NY Monsoonal periods when particles are suspended <50% of the time [Ogston et al., 2008]. During the November, late SE Trades transitional period, the winds are light and variable as the Intertropical Convergence Zone shifts southward and the SE Trade-Wind conditions weaken [Davies, 2004]. During the other annual transitional period in April, winds are again light and variable when the equatorial lows are centered over New Guinea [McAlpine et al., 1983].
 A large, clockwise gyre circulates water in the northern Coral Sea basin [Wyrtki, 1962; Andrews and Clegg, 1989], dispersing sediment and terrestrial and mangrove-derived organic debris predominantly northward and eastward along the coastline [Ogston et al., 2008]. This water mixes with and becomes diluted with the fluvial output of other rivers flowing into the Gulf [Wolankski and Eagle, 1991; Brunskill et al., 1995; Wolanski et al., 1995]. During energetic SE Trade-Wind conditions, surface gravity waves and wind-driven circulation intensify. They couple with strong tidal currents, which vary in strength fortnightly, resuspending sediment and moving it in part as fluidized mud (near-bed suspensions having sediment concentrations 10 g l−1 to ≫100 g l−1) shoreward of the ∼15 m isobath [Walsh et al., 2004; Ogston et al., 2008]. During periods of NW Monsoonal-Winds this fluid or mobile mud is temporarily deposited within the topset zone (∼15–25 m) and stabilized for short (days to weeks) or long (weeks to months) periods of time [Wolanski and Eagle, 1991; Wolanski et al., 1995; Dalrymple et al., 2002; Harris et al., 2004; Martin et al., 2008]. Periodically, and presumably largely during the period of SE Trade-Winds, these muds are refluidized, mobilized, and carried seaward into the foreset region (generally ∼40 to 60 m) where they accumulate [Walsh et al., 2004]. Thus, the topset and forest regions of the clinoform deposits are separated by a critical boundary between high and low shear stress on the seabed due to combined wave and tidal conditions. As described by Walsh et al. , this boundary or rollover point occurs between 25 and 40 m depending on location.
3. Sampling and Methods
 Forty-four sets of triplicate samples at each bottom station were collected using box, gravity, kasten, or multicorers along several inshore–offshore transects on research cruises during NW-Monsoon (December to March) periods (January 1999, February 2000, February 2003, January 2004), SE Trades (late May to late October) periods (August/September 2003, November 2003), and the NW Monsoon to SE Trades Transition (March to late May) periods (May 2004). Those samples collected during 2003–2004 represent new seasonally continuous data; those from 1999–2000 were published previously [Aller and Aller, 2004]. The biological data of Alongi and colleagues [Alongi, 1991; Alongi et al., 1992; Alongi, 1995] have been similarly organized by season of collection. The locations of the stations occupied on the various expeditions are indicated in Figure 1.
 Salinity and temperature profiles were obtained with CTD casts made at each station using a rosette of 10-L Niskin bottles controlled by a Seabird CTD fitted with a PAR sensor, transmissometer, oxygen sensor, and fluorometer. Water samples were taken on the upcast at predetermined depths after evaluating water column structure. Surface and bottom water samples for bacteria and viruses as virus-like particle (VLP) microscopic counts and pigment analyses were collected in triplicate at each station directly from the Niskin bottles (bottom water ∼1 m above seabed). Samples for counts were preserved with formalin to a final concentration of 3%, and kept refrigerated until processed. Water for pigments was filtered onto 25 mm GF/F filters, which were then frozen (−10°C) and stored until subsequent grinding and extraction in 90% acetone and analysis by high performance liquid chromatography (HPLC).
3.1. Macroparticulate Detritus, Particulate Organic Carbon, and Photopigments
 Particulate organic detritus fractions (>1 mm, 1–0.5 mm, and 0.5–0.3 mm) originally retained with sieved macrofaunal samples were recollected after animals were picked from triplicate samples. Detritus was treated overnight with 1 N HCl at room temperature to remove carbonates, washed and dried at 60°C to a constant weight.
 Particulate organic carbon (POC) was determined on bulk sediment and suspended particles as the difference between total C and C after removal of inorganic carbonates with 10% HCl. Weights of dried suspended matter were collected by filtering a known volume of surface water through precombusted and weighed Whatman GF/F glass fiber filters. Filters and sediment samples were placed in tin boats and analyzed with a Carlo Erba 1108 CHNS analyzer. The relative precision was 2%.
 Chlorophyll a, phaeophytin a, and total phaeophytin concentrations were estimated in 100% acetone extracts (24 h, 4°C, dark) of suspended matter from known volumes of seawater or ∼1 g of previously frozen, wet sediment, following the processing method of Green et al. . Chl-a and Pha-a were determined in 50 μL aliquots of extract by ion-pairing reverse-phase high performance liquid chromatography (HPLC) [Mantoura and Llewellyn, 1983; Sun et al., 1991]. The HPLC methods (e.g., eluant concentration, ramp and hold times, wavelengths, and instrumentation) are described in detail in the work of Sun et al. . Identification of Chl-a and Pha-a retention peaks were determined by co-elution of a standard (Sigma Chemical Co.) and quantified spectrophotometrically on a Hewlett Packard 8452 A Diode Array Spectrophotometer using an extinction coefficient of 68,700 at 440 nm. Concentrations are based on integrated peak areas.
 Chl-a analyses by HPLC allow correction for humic interferences and yield higher values than the more common fluorometric measurements (unpublished data) used by Alongi and colleagues. Therefore, for comparison with their studies, total phaeophytin concentrations were estimated fluorometrically on an Hitachi F-4500 Fluorescence Spectrophotometer using optimized wavelengths based on standards of 410 nm excitation and 467 nm emission (predominantly Pha-a) after acidification with 0.1N HCl (wavelengths optimized based on standards) [e.g., Boto and Bunt, 1978].
3.2. Bacteria and Virus Analyses
 Preserved water column samples were filtered for bacteria and virus-like particle counts (VLP). Sediment core subsamples for bacteria counts were taken from multiple depth intervals at increasing thickness with depth: 0–1, 1–2, 2–3, 3–5, 5–10, 10–15, 15–25 cm and subsequently 10 cm intervals to the base of the core (up to ∼2 m). Data are reported for different intervals for comparison with other measurements, which were not available for all depth intervals.
 Bacteria and VLP counts for each sample were made on duplicate slides using a 100× Nikon Eclipse E400 epifluorescence microscope with a MetaMorph™ Imaging System. A minimum of 100 cells per slide were counted for each sample, giving an estimated precision better than 5% (relative standard deviation). Bacterial abundances were determined after staining with the fluorescent dye 4′, 6-diamidino-2-phenylindole (DAPI) after Porter and Feig . VLP counts were made using SYBR Gold, a nucleic acid stain that detects double- or single-stranded DNA or RNA (Molecular Probes, Inc.), and a modification of published procedures [Hewson et al., 2001]. Slides were generally prepared immediately upon return of samples to the laboratory; however, this was not always possible. The effect of sample storage was experimentally examined. Counts of VLPs in preserved water samples were found not to vary significantly (P > 0.05) if counted within 2 weeks of sample collection. There were no significant changes (P > 0.05) in counts between slides prepared and counted immediately after sampling and ones prepared immediately but held in the dark and refrigerated for more than 6 weeks before counting. Counts of bacteria also showed no significant changes (P > 0.05) over a 4-week time period in refrigerated, preserved water and sediment samples. Estimates of bacterial biomass were calculated from average carbon contents determined for bacteria in coastal environments by Fukuda et al.  of 30.2 ± 12.3 fg C cell−1.
3.3. Macrofaunal Abundances and Biomass Estimates
 Seabed sediments were sampled for macrofauna to depths of 25–30 cm using either Smith-McIntyre grabs: 0.1 m2 surface area, 15-cm diameter gravity cores (0.018 m2), box cores (20 × 30 cm cross section), a stainless steel kasten coring unit with a 3 m × 12.5 cm × 12.5 cm barrel [Kuehl et al., 1985] as modified by Brunskill et al. , or in a few cases, a multicorer with a 10 cm diameter core (0.008 m2). Cores were subsampled on the ship either immediately or within a few hours of collection. Although a range of areas were sampled, in all cases the macrofaunal and other biological measurements were made over seabed areas that were either equivalent to or substantially exceeded the areal coverage of physical and chemical measurements at the same sites and to which they are compared. In addition to quantitative faunal sampling, qualitative observations of all bottom samples (e.g., those used for X-radiography, radiochemical analyses, etc.) obtained from a given site were also made to ensure that samples processed for biological components were visually representative and that the presence of any rare species or individuals (e.g., large) were not entirely missed.
 Three separate bottom samples were collected for faunal analyses at each station. The same samples were used for determination of macroparticulate organic detritus. Water overlying the grabs/cores was siphoned into a bucket and sieved along with the top 5 cm of sediment. Grabs/cores for macrofauna were horizontally sliced into intervals at 0–5 cm, 5–10, and 10–25/30 cm. Each depth interval was gently washed with filtered seawater through a nest of sieves having 1 mm, 0.5 mm, and 0.3 mm meshes. Retained faunal and macrodetritus samples were then preserved with 10% buffered paraformaldehyde, stained with rose bengal, and kept cool until processed. Macrofauna were subsequently picked from each sieved sample under a dissecting microscope, sorted to major taxonomic groups, and enumerated. Biomass in wet weight (to 0.1 mg accuracy) for each major taxon was estimated after drying the animals a few seconds on absorbent paper. Biomass values were converted into organic C per major taxon using conversion factors given by Rowe [1983; see Aller and Aller, 2004].
 Unless otherwise stated, all macrofaunal and other biological measurements are expressed as means with standard deviations based on variations between at least triplicate samples. Statistical significance of differences between sample means was determined, when warranted, using ANOVA and t-tests.
4.1. Water Column Properties
 Major features of the water temperature and salinity distributions are described here for a subset of central Gulf stations on the 2003 and 2004 cruises in order to demonstrate both the magnitudes of these properties and their overall seasonality across the system during the study (Table 1). The water column was well oxygenated (>70% saturation) at all times [Aller et al., 2008]. More detailed temperature and salinity data and interpretive analysis for the proximal Fly Delta and additional central Gulf sites can be found in the work of Wolanski et al. , Robertson et al. , McKinnon et al. , and Ogston et al. . Temperatures ranged from 26.4 to 31.6°C over the study area. Distinct across-shelf and vertical patterns in the water column were observed (Figure 2a). The seaward regions (foreset, bottomset) showed a well-defined vertical structure with warmer surface waters, cooler bottom waters (∼2.4 ± 0.8°C difference), and relatively subdued seasonality. In contrast, the topset and inner-foreset zones experienced significant seasonal variations. Water was warmer overall during the NW Monsoons (∼28–30°C) than SE Trades (∼26–28°C), and of intermediate temperature during the Transition period (May). With the exception of the inshore mangrove channel site GH1, vertical temperature gradients over the topset were more pronounced during the NW Monsoon than SE Trades. Topset surface waters tended to be warmer than bottom waters during the NW Monsoon and slightly colder than bottom water during the Transition. At inner-topset stations (water depths <20 m), surface and bottom-water temperatures were essentially uniform during the SE Trades.
Table 1. List of Sampling Locations in Gulf of Papua (Research Vessel)
Water Depth (m)
Latitude (S)/Longitude (E)
HM9901 (R/V Harry Messel); January 1999-NW Monsoon
Inner Topset (mobile muds)
Mid Topset (mobile muds)
FR0002 (R/V Franklin); February 2000-NW Monsoon
Mid Topset (mobile muds)
Mid Outer Topset
CF0103 (R/V Cape Ferguson); February 2003-NW Monsoon
CF0203 (R/V Cape Ferguson); November 2003-SE Trades
Bamu River mouth
E side Purutu Island
Mid Outer Topset
MV0404 (R/V Melville); May 2004-Transition
Proximal Fly Inner Topset
Inner Topset (mobile muds)
Inner Topset (mobile muds)
 Salinity profiles show that in the most inshore/shallowest regions (<10 m), bottom-water salinities were ∼18.5–22, becoming relatively saline (>31) compared to surface waters at bathymetric depths >15 m, and remaining roughly constant seaward of the outer-topset (Figure 2b). Small decreases in bottom-water salinities (∼1–2), however, are evident during the SE Trades and Transition. In contrast, surface waters have marked seasonal variation in salinities with a general decrease in surface salinities during the 2003 SE Trades relative to either the 2003 or 2004 NW Monsoon periods. There are no major differences in surface salinity patterns between the 2003 and 2004 NW Monsoons. Because the 2003 NW Monsoon period was impacted by El Niño conditions (i.e., low river runoff) and 2004 was not [Ogston et al., 2008], the relative NW Monsoon-SE Trades differences observed in the central Gulf are likely robust. These seasonal salinity patterns are also consistent with model calculations of Wolanski et al. . In general, the topset region, which represents ∼70% of the clinoform surface area, experiences substantial seasonality in water column properties, and is relatively warmer and more saline during the NW Monsoon compared to the SE Trades or NW Monsoon → SE Trades Transition periods.
4.2. Chloropigments and Macroparticulate Debris
 Surface water Chl-a concentrations, a photopigment biomarker representing the most labile fraction of recently produced organic material, were usually more than twice those of bottom water concentrations, although local exceptions occur (Figure 3). The decrease in Chl-a with depth in the water column most likely reflects lower primary production and is consistent with an overall increase in turbidity and decrease in light with depth (PAR data not shown [Davies, 2004; McKinnon et al., 2007; Ogston et al., 2008]), although zooplankton grazing pressure (not measured) might have had an impact. The highest surface Chl-a concentrations across the system, and the greatest differences between surface and bottom water Chl-a values, occurred during the Transition period [McKinnon et al., 2007], and the lowest values in both surface and bottom were found during the NW Monsoons (Figures 3a and 3b). During both the NW Monsoon and SE Trades, there was an inner-topset to foreset decrease in Chl-a, consistent with general inshore–offshore patterns found in previous studies [Robertson et al., 1998; Davies, 2004]. Seasonally, Chl-a averaged 1.06 ± 0.69 μg L−1 during the Transition period, significantly greater (P < 0.05) than the NW Monsoonal concentration of 0.4 ± 0.4 μg L−1 and the concentration during the SE Trades of 0.47 ± 0.4 μg L−1 in surface waters. Bottom waters across the Gulf were relatively constant (not significantly different P > 0.05) at all seasons, averaging 0.20 ± 0.18 μg L−1 during the Transition period, 0.38 ± 0.28 μg L−1 during NW Monsoons and 0.40 ± 0.33 μg L−1 during SE Trades. Concentrations of Chl-a and phaeopigments in surface sediments (top 1 cm) also show pronounced seasonal differences (Figures 3c and 3d). In contrast to the water column, Chl-a and phaeopigment concentrations were significantly greater (P < 0.05) during the NW Monsoon than during the Transition period, with NW Monsoon Chl-a values averaging 0.62 ± 0.56 μg g−1 dry sediment and pheopigments, 0.77 ± 0.33 μg g−1 (not including the GH1- Wame River site, Figure 1). During the SE Trades periods, Chl-a values ranged from an average of 0.3 ± 0.14 in August/September to 0.58 ± 0.32 μg g−1 in November. This was significantly different (P < 0.05) from concentrations during the NW Monsoon-SE Trades Transition period when they averaged 0.016 ± 0.019 μg g−1 in the central Gulf (>5 m), but not significantly different (P > 0.05) from NW Monsoonal periods. Phaeopigments averaged 0.54 ± 0.32 (August/September) to 0.86 ± 0.74 (November) μg g−1 during Trade periods, significantly (P < 0.05) different from NW Monsoonal periods but not different (P > 0.05) from Transition periods when they averaged 0.97 ± 0.91 μg g−1. In contrast with central Gulf sediments, both Chl-a and phaeopigments remained elevated at all seasons in the water column and surface sediments within the delta plain channels of the Fly River (Alongi et al. ; data not plotted).
 Total particulate organic carbon concentrations (mmol C g−1) in suspended matter (SPM POC) varied seasonally in surface water but, with the exception of enhanced values at proximal Fly River sites during the 2003 NW Monsoon [Goñi et al., 2006], showed relatively little across-shelf variation in the central Gulf clinoform during any particular season (Figure 4a). Surface water SPM POC was generally 2–3X greater than bottom water at all times. The highest seasonal values of surface water SPM POC were found during the SE Trades. Differences in the relative concentrations of Chl-a and POC also occur seasonally, with substantially higher values of POC compared to Chl-a common during the SE Trades (median POC/Chl-a = 254 mg C μg Chl-a−1) than NW Monsoons (median POC/Chl-a = 131 mg C μg Chl-a−1). These Chl-a - POC relations imply that the enhanced water column SPM POC present during the SE Trades reflects input of terrestrial detritus such as mangrove debris depleted in the most recently formed labile plankton components.
 Concentrations of macroparticulate plant debris (>0.062 mm) in the upper 0–1 cm of bottom sediments, were significantly greater (P < 0.05) during the NW Monsoons than during either the SE Trades or Transition periods (Figure 5). Inshore of 30 m, concentrations averaged 0.16 ± 0.13 g g−1 dry sediment during the NW Monsoons compared with 0.035 ± 0.024 during the SE Trades and 0.03 ± 0.025 g g−1 dry sediment during the NW Monsoon → SE Trades Transition period. The monsoonal seafloor surface maximum in detrital debris follows the SE Trades seasonal maximum in the water column (Figures 4a and 5).
 The seafloor inventory of macroparticulate plant detritus over the upper 30 cm of sediment, which is presumably a more time integrated measure than the upper 1 cm alone, also varied both spatially and seasonally. Inventories initially decreased approximately exponentially offshore from inshore mangrove channels across the mobile mud zone and inner-topset stations. Secondary maxima concentrations occurred at outer-topset stations in water depths of ∼25–30 m depth, and values steadily decreased thereafter to <700 g m−2 at water depths of 75 m within the bottomset. The relative shelf-wide across-shelf distribution was retained at all seasons (Figure 6); however, inventories in the upper 5 cm in particular appeared to vary seasonally at many sites. The highest inventories of macrodetritus were found during NW Monsoons and Transition periods (seaward of ∼15 m; no topset sites <15 m available). Concentrations >3000 g m−2 in the top 0–5 cm and >5000 g m−2 over depths of 5–30 cm were found at stations <5 m in the river mouths of the Bamu and Pai-a Rivers. Inventories of macrodetrital debris reached 5300 ± 390 g m−2 at station GH25 in 25 m of water during the Transition period.
4.3. Water Column Bacteria and Virus-Like Particles (VLPs)
 Although abundances of bacteria in the water column do not show strong seasonal or spatial patterns across the region, in the present study (2003–2004), slightly higher abundances of both bacteria and VLPs were found at inshore sites and over the topset, decreasing offshore (Figure 7); however, this difference was not statistically significant by ANOVA test criteria (P < 0.05). The seasonally greatest abundances of water column bacteria occurred during the SE Trades which were significantly higher (P < 0.05) than during NW Monsoon periods, although not significantly different (P > 0.05) from abundances during the Transition. VLP abundances were significantly lower (P < 0.05) during the NW Monsoons than either SE Trade or Transition periods. Abundances of bacteria and VLPs showed no regular difference between surface and bottom waters, although there is an indication during the Transition period of higher concentrations in bottom than surface waters at several sites (Figures 7a and 7b). Data collected during the present study show a rough positive correlation between bacteria and viruses, with an elevated clustering of surface and bottom water samples during the SE Trades at times of highest SPM detritus concentrations (Figures 7c and 4), presumably reflecting the active replication of viruses in growing bacteria.
4.4. Benthic Bacteria and Macrofauna
 The distributions of bacterial abundance in the surfacemost sediment layer (0–1 cm), which should be responsive to boundary conditions, showed no regular water depth dependence across the Gulf but had distinct seasonal patterns. Elevated numbers were most common during NW Monsoon periods (1999–2004) and relatively low abundances occurred during Transition and SE Trades (Figure 8). NW Monsoonal periods also showed the greatest degree of spatially variability with concentrations varying from 1.3 to 7 × 109 cells g−1. The averaged abundances of both benthic and water column bacteria correlate directly with Chl-a concentrations, reflecting the mirror image seasonal availability of labile organic components in the two reservoirs (Figure 9).
 The inventories of macrobenthos in the top 30 cm reflect seasonal as well as spatial differences in the Gulf (Figure 10). The most obvious spatial feature in the central Gulf is the relative depression of macrofaunal numbers and biomass on the topset between the 5–20 m depth range at all times (Figures 10 and 11), a region representing ∼70% of the surface area of the central Gulf clinoform. When high abundances were observed on the inner topset, the fauna were composed of very small individuals (0.5–0.3 mm), large individuals were virtually absent and biomass was minimal. Macroinfaunal populations are also depauperate in the migrating channel deposits of the proximal Fly River delta plain (0–5 m water depths) [Alongi, 1991]. The low abundance of macroinfauna in these regions <20 m water depth is reflected in the dominance of physical sedimentary structures preserved (example x-radiographs in the work of Alongi , Aller and Aller , Walsh et al. , Goñi et al. , and Aller et al. ). Where biogenic sedimentary structures are evident in x-radiographs obtained from the topset, for example, they are often truncated by erosive horizons or buried by depositional layers, indicative of failed colonization.
 Macroinfaunal abundances and biomass increase seaward by 10–100X at depths of 25 to 35 m (outer topset, rollover region). Macrofaunal biomass is, however, always subordinate to bacterial biomass and small compared to most other shelf systems [Aller and Aller, 2004, Figure 11]. Relatively large individuals (>0.5 mm) almost always dominate the macrobenthos at the deeper sites, and sediment fabric shows increasing biogenic structure. The sedimentary structure patterns hold along depositional strike [Walsh et al., 2004], implying spatially similar faunal distributions. On the foreset >50 m, there was little seasonal difference (2003–2004) although macrofaunal abundances tended to be slightly lower during NW Monsoon periods compared with other seasons. The greatest seasonal variability in abundances (but not biomass) observed at any site in the present study occurred in the Umuda Valley, a sediment conduit off the Fly River characterized by reworked or mobile deposits in the upper ∼1–1.5 m [Martin et al., 2008]. Interannual differences in abundances in the central Gulf are apparent when data collected by the same kind of sampling gear and using the same methodology during the 1999 NW Monsoon, are contrasted with data collected during the 2003 and 2004 NW Monsoons (Figure 10), the latter were generally an order of magnitude greater within the foreset (the topset is always depauperate).
 The present seasonal study of benthic community patterns in the Gulf of Papua confirms the overall spatial distributions and the relative community compositions documented previously. Our study further resolves likely structuring factors related to seasonal coupling between water column and benthic processes across the clinoform, and, taken together with past studies, provides a decadal perspective on the stability of the system.
5.1. Spatial and Temporal Patterns in Benthic Communities
 Past and present studies demonstrate that Gulf benthic communities are characterized by low abundances and biomass of macroinfauna and show that macrofauna are particularly depauperate at all times on the muddy topset between ∼5 and 20 m depth (Figures 10–12). The relative patterns in community structure (bacterial dominance) appear to be stable over decadal timescales although absolute abundances and biomass vary seasonally and interannually. Abundances and biomass of macrofauna in the topset depth range are commonly 10–100X lower than in either shallow inshore mangrove regions or in the deeper foreset. In fact, the complete absence of macrofauna is not unusual in topset samples. This zone comprises ∼70% of the area of the clinoform, thus from an areal standpoint a large proportion of the system lacks significant macrobenthic populations. When they do occur in the muddy topset deposits, macrofaunal communities are frequently composed of small deposit feeders or tubiculous amphipods indicative of early colonization stages of disturbed habitats.
 In contrast, bacterial communities appear diverse, active, and dominate benthic biomass to bathymetric depths >60 m [Alongi, 1995; Alongi and Robertson, 1995; Todorov et al., 2000; Aller and Aller, 2004, Figure 11]. Bacterial biomass within the various depositional regions, including the topset, is comparable to the highest values reported for continental shelves [Aller and Aller, 2004]. Although macrobenthos increase in abundance, biomass, and size in the foreset zone, and their activities progressively impact sedimentary structure within the upper 20 cm, their biomass remains subordinate to bacteria by factors of 10–100X throughout the system and low compared to most other shelves (Figures 10 and 11). These overall spatial patterns appear predominantly to reflect inhospitable physical sedimentation conditions and frequent mobilization and reworking of topset deposits, particularly during the SE Trades.
 The spatial pattern of relatively few macroinfauna over the topset is temporally stable; however, annual and interannual variation in absolute abundances can be substantial, particularly in the rollover region and the foreset (Figure 11). In the present study, it is not possible to discern consistent seasonal patterns across the foreset, although macrobenthic abundances and biomass were seasonally greatest in the rollover–upper foreset zone (25–40 m) during the NW Monsoons and least during the SE Trades. Far greater relative variations occurred interannually between the 1999–2000 and 2003–2004 Monsoons, with the latter period typically having 10–50X greater abundances and biomass within the foreset than the earlier sampling. Macrofaunal abundances within the foreset comparable to those in the 2003–2004 NW Monsoons were also found at a few sites for the 1990 NW Monsoon period by Alongi et al. , indicating that significant (P < 0.05 by ANOVA) interannual variations are not unusual in the foreset region. We believe the inter-NW Monsoonal and decadal differences in the deeper regions are related primarily to variation in food resources, as outlined subsequently.
5.2. Benthic Food Resources
 In addition to physical disturbance, the availability of food likely plays a role in determining both the spatial and seasonal distributions of benthos in the Gulf [Alongi and Robertson, 1995]. Benthic food resources consist of inputs of labile planktonic debris from water column primary production, local biomass, terrestrial macrodetritus, and the more refractory organic matter associated with mineral particles. These inputs clearly can change seasonally and spatially (Figures 3–6). For example, exceptionally high values of organic detritus in inshore surface waters reflect proximity to mangrove islands and the tremendous local flux of plant debris comprised mainly of macerated Nypa palm fronds and mangrove litter [Robertson et al., 1998; Goñi et al., 2006]. The relative increase of Chl-a, organic detritus, water column microbes, and decreases in salinity and temperature over the topset observed during the SE Trades in the present study implies seasonally elevated water column primary production as well as enhanced delivery of terrestrial detrital substrates to the Gulf during the SE Trades compared to the NW Monsoon period. Although most previous water column studies have emphasized either the inshore estuarine channels or offshore zones (foreset), past observations are also generally consistent with relatively increased water column Chl-a, bacterial abundances, suspended matter, and primary production, and with lowered salinities (greater river influence) over the topset during the SE Trades [Wolanski et al., 1995; Robertson et al., 1998; Davies, 2004].
 In terms of indicators of food availability and production, the seabed apparently follows a mirror image seasonal pattern to that found in the water column (Figures 3–8). Planktonic debris (Chl-a), macrodetritus, in situ biomass synthesis (bacteria, macrofauna) and benthic remineralization rates [Aller et al., 2008], all of which are indicative of labile substrate availability, increase substantially in surface sediment during the Monsoons relative to the SE Trades or Transition (Figures 3–8 and 13). These effects extend into the outer-topset and to a lesser extent the foreset zone, indicating deposition and export from the shallow topset regions during the latter part of the SE Trades. A seaward decline in both labile and refractory organic matter in the seabed is reflected by a distinct decrease in remineralization rates at most depths >30 m as shown by ΣCO2 production flux measurements (Figure 12) [Aller et al., 2008]. Benthic remineralization rates directly reflect the recent input of labile organic substrates [e.g., Martin and Bender, 1988]. Whereas bacteria respond to the seasonally increased labile substrate throughout the deltaic system, the macrofaunal response is largely found in the foreset where physical conditions are more conducive to survival.
 These various relationships suggest a dual role of local physical disturbance and food availability in governing both the spatial and temporal (both short-term and longer, i.e., decadal) distributions of benthic populations across the deltaic system. The topset is sufficiently physically disturbed that only the bacterial communities respond effectively to changes in substrate supply. In contrast, the deeper foreset region, while subject to episodic input of sediment in restricted areas like drowned river channels areas in the Umuda Valley, is generally less directly impacted by current and wave reworking but is nevertheless dependent on deposition of reactive substrates. The longer-term variations observed in macrofaunal populations on the deeper foreset zone, for example: 1999 NW Monsoon versus 2003–2004 or 1993, presumably reflect increases and decreases in food resources related to primary production and export of detritus from the topset rather than major seasonal or decadal changes in local physical disturbance. The macrofaunal community requires time to respond to food input, and we hypothesize that the particularly low macrofaunal abundances and biomass measured along the foreset in 1999–2000 despite relatively high sedimentary Chl-a, reflect the recent deposition of labile substrate and insufficient time for macrofaunal response. While terrigenous sedimentation on the foreset appears to take place in pulses, which must also input food, it is sufficiently regular in nature that steady accumulation is a good approximation. Major changes in bulk sediment delivery (i.e., catastrophic mass burial) are unlikely to account for interannual or decadal patterns in the foreset benthic communities [e.g., Brunskill et al., 2003; Walsh et al., 2004; Aller et al., 2008].
 Predation and commercial harvesting activities are probably also significant structuring factors, particularly in water deeper than ∼20 m, and must further modify the observed distribution patterns. Commercial prawn fisheries are focused within the outer topset and rollover region (∼20–30 m), however, and there is no obvious evidence of unusual benthic pressure from epifaunal or infaunal predator abundances or distributions across the clinofrom that might account for macrobenthic dynamics or distributions [Alongi et al., 1992]. Such biological interactions remain to be more fully constrained.
5.3. A Conceptual Model of Benthic Dynamics
 Strong seasonal and spatial differences in physical conditions, biological community characteristics, and biogeochemical cycling regimes are apparent in the Gulf of Papua water column and seabed, particularly across the topset (Table 2). The seasonal dynamics of the benthic system, based on observed spatial and temporal patterns of labile/refractory organic matter pools in the water column and seabed, the benthic community characteristics, and net benthic remineralization rates are summarized by the conceptual model depicted in Figure 14. The SE Trades are a period of intense disturbance and reworking of the shallow seabed, increased terrestrial organic matter inputs from the shoreline into the Gulf, and stimulated water column primary production, apparently with minor net deposition of biologically labile substrates to the bottom. In contrast, the NW Monsoons represent a period of net deposition of reactive material (planktonic, macrodetritus) to the seabed, presumably early in the season, and enhanced benthic remineralization. Possible decadal or interannual differences in primary production and net deposition of reactive material are reflected in bacterial production as evidenced by greater abundances during the NW Monsoonal period of 1993 reported by Alongi and Robertson  and the 1999, 2003, and 2004 data presented here. Export of relatively labile substrates to the deeper foreset promotes macrobenthic community development. As indicated by sedimentary structures and biological samples, macrobenthos also presumably attempt recolonization of the topset during this period of relative quiescence and favorable food resources. Reduced riverine sediment input and lack of major depositional events, perhaps during El Niño years, may also temporarily promote macrobenthic community development. During the NW Monsoon → SE Trades transition, initial resuspension of nutrient-rich sediments and water column mixing stimulate primary production, perturb, reduce or annihilate benthos within the topset, and begin the next cycle. We present this model as an overall hypothesis consistent with existing data and as a guide to future studies. More continuous seasonal coverage and integrated water column and seabed data across the system are required to further test the model.
Table 2. Comparison of Seasonal Variations in Water Column and Seabed Characteristicsa
Terms are relative. Exact details can be found in the text.
Highest overall topset temperatures
High Chl-a (0–1 cm)
Highest topset salinities
High macrodetritus (0–1 cm)
Topset/foreset surface warmer than bottom
High abundances of bacteria (0–1 cm)
Lowest Chl-a surface values
High abundances of macrobenthos
Lowest abundances of bacteria
High remineralization rates
Greatest bacterial and macrofaunal biomass (<40 m)
Highest Chl-a surface water values
Low Chl-a (0–1 cm)
Intermediate abundances of bacteria
Low abundances of bacteria
Low remineralization rates
Coldest topset temperatures
Low to intermediate Chl-a (0–1 cm)
Lowest topset salinities surface, bottom
Lowest sediment macrodetritus
High Chl-a surface values
Fewest macrobenthos (<40 m)
Highest POC, POC/Chl-a
Low remineralization rates
Highest abundances of bacteria
Elevated VLP abundances
 We are very grateful to Gregg Brunskill, Australian Institute of Marine Science, whose coordination and support of project TROPICS made this research possible. The expertise and efforts of the Masters and crews of the R/V Harry Messel, R/V Franklin, and R/V Cape Ferguson (all from AIMS) and R/V Melville (Scripps Institute of Oceanography) were important to the success of the project. Vanessa Madrid, Irena Zagorskis, John Pfitzner, Gregg Brunskill, Caterina Panzeca, Megan Dantzler, Lynn Abramson, Vasso (Bessy) Alexandratos, Miguel Goñi, and Chuck Nittrouer provided invaluable aid in the field. Thank you to David McKinnon for water column productivity, Chl-a, and CTD data from the Transition period. C. Heilbrun and L. Abramson carried out HPLC pigment analyses. The constructive comments on this manuscript from P. Jumars, C. Nittrouer, and an anonymous reviewer are appreciated. Financial support came from NSF Ocean Sciences Program, OCE 9818574.