Lignin, elemental, and stable carbon isotope compositions are reported for local plants and for coarse (>63 pm) and fine (~63 pm) suspended particulate materials collected along a 1,950-km reach of the lower Amazon River during four contrasting stages of the 1982-1983 hydrograph.

Sediment trap and net plankton samples were collected monthly for a year at three depths in a marine bay (Dabob Bay, Washington). These materials and subsamples from a sediment box core were analyzed for lignin oxidation products as well as elemental and stable carbon isotope compositions. The sediment core was compositionally uniform over its entire 50-cm length. The elemental and lignin compositions of the sediment trap and core samples indicate nitrogen-rich (atomic C : N = 7.5) plankton-derived organic matter mixed with vascular plant debris.At most, vascular plant debris accounts for 10% (nonwinter months) to 35% (winter months) of the total organic carbon in the upper water column (30 m) sediment trap samples and consists predominantly of gymnosperm wood along with some nonwoody gymnosperm tissues and angiosperm woods. Bulk land-derived organic matter in Dabob Bay contains a maximum of 50% vascular plant debris and comprises an average of one-third of the total organic carbon in the sediment trap samples and two-thirds of the total organic carbon in the underlying sediments. Lignin in the sediment trap and core samples shows evidence (from elevated vanillic acid:vanillin ratios) of white-rot fungal degradation before (but not after) introduction to the water column at the study site. Vascular plant debris introduced to the bay has already lost almost half of its initial bulk polysaccharide. Glucose yields are particularly low whereas rhamnose and fucose are obtained in excess of expected yields and must have additional sources.Lignin and neutral sugars together account for ~20% of the total organic carbon in the sediment trap and core samples. Overall, the sediments of Dabob Bay compositionally resemble the gymnosperm wood-rich particulate material introduced to the overlying water column during winter and poorly record the input of plankton and other types of vascular plant debris during nonwinter months.

Methane emissions to the troposphere from the Amazon River floodplain have been determined. In total, 94 individual flux measurements were made along a 1,700-km stretch of the river during the early falling water period of the flood cycle (July-August 1985). The overall average rate of CH 4 emission from wetlands was 390 mg CH4 m -2 d-x. Water surfaces covered by aquatic macrophytes had the highest emissions, 590 mg CH½ m -2 d -•, while emissions from flooded forests were 110 mg CH,• m -2 d -• and open lake areas averaged 120 mg CH½ m-2 d-•. Ebullition was the dominant mechanism of emission, accounting for 85% of the total. Surface water dissolved oxygen was always undersaturated, while dissolved CO2 concentrations were significantly supersaturated with respect to atmospheric equilibrium. Surface water CH½ concentrations were highly supersaturated, averaging 6.4 pM. The g•3C of the CH½ varied from -42%0 to -73%0 with an apparent trend toward heavier values downstream. Extrapolation to the entire Amazon floodplain results in estimated annual emissions in the troposphere of 10 Tg yr-• and indicates the importance of the area in the global atmospheric CH4cycle. Recent evidence indicates that the atmospheric concentration of methane has been increasing at a rate of between 1-1.5% per year [Rasmussen and Khalil, 1986; Khalil and Rasmussen, 1986; Pearman et al., 1986; Stauffer et al., 1985]. As CH,• is an important "greenhouse" gas and plays a role in the atmospheric cycles of O 2, C1, and OH radical [Crutzen, 1973; Wofsy, 1978; Logan et al., 1981], this observed increase could be of significant environmental consequence. Current global CH,• budgets place inputs to the atmosphere at between 370 and 1000 Tg yr-•, with the majority of the input being biogenic in origin [Ehhalt and Schmidt, 1978; Khalil and Rasmussen, 1983; Sheppard et al., 1982; Seiler, 1984]. Although estimates of the magnitude of the various components of the biogenic emissions vary between budgets, CH,• emissions from wetlands are generally thought to account for 20 to 40% of the total. These estimates, however, are based on only a few field measurements, and virtually no data are available from the extensive wetland areas of the tropics, e.g., the Sudd in Sudan and the Amazon River floodplain and Pantanal regions of Brazil. Here we present the results of a study of mechanisms and magnitudes of CH,• emissions to the troposphere from the Amazon River floodplain. During its annual flood cycle the Amazon River rises 10 m, inundating its approximately 100,000 km 2 floodplain or "varzea." Depending on the flood stage, during a normal year between 40 and 80% of the floodplain is covered by shallow lakes, swamps, and dense beds of floating aquatic macrophytes [Junk, 1985]. As the varzea is a region of high primary production and uniformly warm temperatures, these wetlands constitute a significant potential for CH,• production [Richey et al., ment and as part of the Carbon in the Amazon River Experiment (CAMREX) project, during the early falling water period ...

Vertical fluxes of bulk particulate material, organic carbon, nitrogen, lignin-derived phenols, and neutral sugars through the water column and into surface sediments of Dabob Bay, Washington, were determined monthly for 1 yr by sediment trap deployments at 30, 60, and 90 m at a site 110 m deep. Vertical fluxes of sinking bulk particulate material in this marine bay were elevated during winter and increased in consistent proportion to sediment trap deployment depth throughout the year. Although annual average particle fluxes at 30 and 60 m bracketed the mean accumulation rate of the underlying sediment, the flux at 90 m was higher by a factor of 2 due to resuspension, horizontal advection, or both.The monthly fluxes of lignin-derived phenols paralleled those of total particulate material, indicating a common riverine origin. The annual average fluxes of vanillyl and cinnamyl phenols through the water column closely matched the corresponding accumulation rates in the underlying sediment, whereas about a third of the total syringyl phenol input was degraded at the watersediment interface. Although p-hydroxyacetophenone exhibited a stability typical of lignin-derived phenols, the distinctly higher reactivities (> 60% degradation) ofp-hydroxybenzaldehyde and p-hydroxybenzoic acid indicate a predominantly nonlignin source.On average, 60 and 70%, respectively, of the total particulate organic carbon and nitrogen and 65-75% of all neutral sugars settling through the midwater column were degraded at the watersediment interface. The elemental and carbohydrate composition of the degraded material was similar to that of local net plankton except for higher percentages of glucose and total neutral sugars. Land-derived organic material accounted for about one-third of the total organic carbon passing through the midwater column and two-thirds of the organic carbon accumulating in the underlying sediments. The amounts of plankton-derived organic matter sinking through the midwater column and being preserved in the sediments below corresponded to 14 and 3% of the annual mean primary productivity. Plankton-derived organic matter exhibited about 5 times the reactivity of local land-derived organic matter at the water-sediment interface of Dabob Bay and supported essentially all of the benthic respiration.